I and AI: Human Personhood in an Age of Artificial Intelligence
I and AI: Human Personhood in an Age of Artificial Intelligence
BY
Jonathan A. Powers & Daniel H. Strait, eds.
ISBN: 9781648174353
I and AI : human personhood in an age of artificial intelligence
Edited by Jonathan A. Powers & Daniel H. Strait
Published in the U.S.A. by First Fruits Press, 2026. Digital version at http://place.asburyseminary.edu/firstfruitsacademicbook/
First Fruits Press is a digital imprint of the Asbury Theological Seminary, B.L. Fisher Library. Asbury Theological Seminary is the legal owner of the material previously published by the Pentecostal Publishing Co. and reserves the right to release new editions of this material as well as new material produced by Asbury Theological Seminary. Its publications are available for noncommercial and educational uses, such as research, teaching and private study. First Fruits Press has licensed the digital version of this work under the Creative Commons Attribution Noncommercial 3 0 United States License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc/3.0/us/
For all other uses, contact: First Fruits Press
B.L. Fisher Library
Asbury Theological Seminary 204 N. Lexington Ave. Wilmore, KY 40390 http://place.asburyseminary.edu/firstfruits
1 Artificial intelligence Religious aspects--Christianity. 2 Artificial intelligence Moral and ethical aspects. 3. Self Religious aspects--Christianity. 4. Persons Religious aspects. 5. Identity (Philosophical concept)
6. Philosophical anthropology Religious aspects Christianity.
I. Powers, Jonathan A. II. Strait, Daniel H.
BR115.T42 I26 2026eb 253.0285/63
Cover design by Amanda Kessinger
First Fruits Press
Te Academic Open Press of Asbury Teological Seminary 204 N. Lexington Ave., Wilmore, KY 40390 859-858-2236
Bill T. Arnold (Ph.D., Hebrew Union College) is the Paul S. Amos Professor of Old Testament Interpretation at Asbury Theological Seminary. His research encompasses Old Testament and ancient Near Eastern studies, both separately and in combination. He specializes in Pentateuchal interpretation and has written on many aspects of Old Testament interpretation, including Hebrew language and the history of Israelite religion.
Kevin J. Brown serves as Asbury University’s 18th president. His education includes an undergraduate degree and an MBA from the University of Indianapolis along with two degrees from institutions in Scotland: an MA of Letters from St. Andrews University and a PhD from the University of Glasgow. Dr. Brown’s graduate background incorporates economics, philosophy, and theology, and his research seeks to explore the interplay between these fields.
K. Paul Nesselroade, Jr. earned an M.A. and Ph.D. in Experimental Psychology from the University of Louisville. He is currently a Professor of Psychology at Asbury University. He also chairs the Psychology Department, serves as Director of the Honors Program, and is the Associate Dean for the College of Arts, Humanities, and Social Sciences.
Claire Brown Peterson currently resides in Knoxville, TN. Previously, she was Professor of Philosophy at Asbury University. She has published in ethics and philosophy of religion.
Jonathan A. Powers earned a Doctorate in Worship Studies from the Robert E. Webber Institute for Worship Studies and a Ph.D. in Worship from Middlesex University in London, England. He serves as the Associate Professor of Worship Studies at Asbury Theological Seminary in Wilmore, KY.
A. Sue Russell is Professor of Mission and Contextual Studies at Asbury Theological Seminary. With doctorates in both New Testament and Anthropology her research brings anthropological insights into theological questions and theological insights into anthropological questions.
Daniel Strait is professor of English and Dean of the College of Arts, Humanities, and Social Sciences. His main research interests include exploring the intersections of literature, philosophy, and religion in Renaissance lyric poetry, especially in the poetry of George Herbert. He has presented and published papers on George Herbert, Shakespeare, Emily Dickinson, Derek Walcott, Robert Browning, G. K. Chesterton, and George Bernard Shaw.
James R. Thobaben (Ph.D., Emory University) serves as the Dean of the School of Theology and Formation and Professor of Bioethics and Social Ethics at Asbury Theological Seminary in Wilmore, KY. In addition to his work at the seminary, Dr. Thobaben is co-pastor at Mt. Zion United Methodist Church in Mercer County, Ky.
Introduction
Kevin Brown, Asbury University
In matters of labor-saving technology, economists differentiate between capital that is complementary and capital that is substitutionary. The former enhances a laborer’s efficacy and output; the latter threatens their replacement. Naturally, this creates an automation anxiety whose industrial roots can be traced back to nine textile workers threatened by the prospects and promises of automated looming.1
In the decades since, these fears have, for the most part, been overstated. Technology has not displaced labor so much as it has elevated men and women to more valuable and humanizing forms of work. Today, “luddite” is a pejorative term implying an irrational fear of technology.
This pattern may have changed, however, on November 30, 2022, with the release of Open AI’s ChatGPT. We have now reached an AI inflection point, the result of which is a variety of other Large Language models (LLMs) promising “super agency” and AI systems surpassing “almost all humans at almost everything.”
Unsurprisingly, anxieties about human relevance in our current AI landscape swell proportionate to the exponential march of technological intelligence. “Will AI Take All Our Jobs?” reads a recent cover of The New Yorker. Derek Thompson, writing in The Atlantic, suggests that graduates are now competing with AI in the labor market.2 And this is to say
1 Automation Anxiety: And a Right to Freedom from Automated Systems and AI, ed. Ziyaad Bhorat (Cambridge, MA: Carr Center for Human Rights Policy, Harvard University, October 2, 2023), Issue 2023-2.
2 Derek Thompson, “Something Alarming Is Happening to the Job Market: A New Sign That AI Is Competing with College Grads,” The Atlantic, April 30, 2025.
nothing of predictions that AI will erode democracy, broadcast misinformation, or even supplant humanity entirely. The range of technological prognostication is wide, offering forecasts that range from Gilded Age optimism to scenarios of apocalyptic doom. A recent editorial in Nature opened with, “It is unusual to see industry leaders talk about the potential lethality of their own product.”3
A common reaction to this anxiety is to argue that humans still exceed AI in matters of reason and intelligence. For example, critics may point out that GPT produces stale and predictable writing, offers up debatable assertions with unfounded authority, or worse, advances inaccuracies. The sentiment behind these critiques is, “Humans still do it better.” This is arguably true. However, the argument itself establishes AI’s value and relevance in terms of efficacy. But if the conditions for AI adoption are distilled to mere considerations of speed and quality, humans will inevitably lose.
At best, this is an unhelpful frame. If the standard for AI’s value is directly correlated to its superiority among alternatives, i.e., “Can technology do this better?”—increasingly the answer will be “yes.” But there are other considerations and standards. What if our understanding of technology, its strategic application, and human relevance was not merely a question of efficacy? What if we begin with a different set of questions—moral, even spiritual, questions?
In the “Techno-Optimist Manifesto,” author and technofuturist Marc Andreessen asserts our anxiety is unfounded. “We are being lied to,” he writes. “We are told that technology takes
3 Stop talking about tomorrow’s AI doomsday when AI poses risks today, editorial, Nature 618 (2023), https://doi.org/10.1038/d41586-023-02094-7. “It is unusual to see industry leaders. talk about the potential lethality of their own product.”
our jobs, reduces our wages, increases inequality, threatens our health, ruins the environment, degrades our society, corrupts our children, impairs our humanity, threatens our future, and is ever on the verge of ruining everything.” In truth, says Andreessen, technology is the “the spearhead of progress.”4 But what, we might ask, is this technological spear aimed toward?
Even on its best day, AI is an apparatus, not a guide; a vehicle, not a map. And yet, it is easy to confuse the instrument for the ends it is meant to actuate. “What I am against,” writes Wendell Berry, “is our slovenly willingness to allow machines and the idea of the machine to prescribe the terms and conditions of the lives of creatures.”5 Berry is rightfully concerned with our willingness to assess technological capacity as our standard for human progress.
In his book The Grand Design, Stephen Hawking reflects on the human capacity to wonder about our world, the universe, the nature of reality, and origins. “Traditionally, these are questions for philosophy, but philosophy is dead,” he says. “Philosophy has not kept up with modern developments in science…Scientists have become the bearers of the torch of discovery in our quest for knowledge.”6
Responding to Hawking’s assertion, Michael Sandel challenges the notion that science and its technological offspring have settled our need for philosophical reflection, discovery, and moral deliberation—that we have somehow graduated from older forms of knowledge now that scientism’s “torch of discovery” has illuminated the world. “[S]cience
4 Marc Andreessen, “The Techno-Optimist Manifesto,” June 2023.
5 Wendell Berry, “Life Is a Miracle” Washington, D.C.: Counterpoint, 2000, p. 54.
6 Stephen Hawking and Leonard Mlodinow, “The Grand Design” (New York: Bantam Books, 2010), p. 5.
and technology, for all their wonders, can’t tell us how they. . . should be used for the human good,” he says. “They can’t tell us . . . the meaning of a just society. They can’t tell us what we owe one another as fellow citizens. These are questions that only we can decide.”7 Among other things, Sandel’s challenge speaks to a more capacious understanding of progress. Sandel reminds us that citizens, not consumers, deliberate questions of truth, justice, beauty, and goodness.
Artificial intelligence is a good servant, but a poor master. It is a spearhead, of sorts, to be sure. But the progress that spearhead is aimed toward will not be determined by the spear, but by its thrower. As a community in possession of such a tool, there is a responsibility to consider the moral and spiritual ramifications of its control and usage—and to bring those considerations to bear against the teleological aims of a created people inhabiting a created world by a loving Creator. This, and nothing else, will be our ultimate standard for human progress.
7 Michael Sandel, "A Night with Michael Sandel," Q&A, ABC, March 26, 2018, https://www.abc.net.au/qanda/a-night-with-michaelsandel/10649342.
Chapter 1
Shaping Minds, Forming Souls: Reflections on AI and Education
Jonathan A. Powers, Asbury Theological Seminary
One of the crucial tasks of educators today is to think critically about artificial intelligence (AI) and its implications for education and theological anthropology. AI presents exciting opportunities and potential benefits, including new efficiencies and innovations that could transform how we teach, learn, and interact in educational settings. Yet, as we embrace these tools, it is important to think critically about the broader consequences of this integration. What might be lost when we prioritize efficiency and innovation above other values? How do we preserve the slow, formative work of education amid the push for innovation?
As we consider these questions, first, it is important to define what “AI” is. Artificial intelligence refers to systems or machines that simulate human intelligence to perform tasks such as decision-making, speech recognition, learning, and problem-solving. AI can analyze vast amounts of data, automate repetitive tasks, and even learn from experience to improve over time. A subset of AI that is particularly relevant today is generative AI, which refers to AI systems capable of creating new content—such as text, images, or even music—based on the patterns it has learned from existing data. Generative AI models, like ChatGPT, can generate written content, answer questions, and engage in creative tasks. This technology offers fascinating potential for innovation in education but also raises important questions about originality, creativity, and the role of human educators.
Reflecting on the continual development of AI, I am reminded of the lessons from Robert Putnam’s Bowling Alone. Putnam’s work demonstrates how various technologies and innovations, while beneficial in some respects, had significant unintended consequences that few anticipated at the time. Air conditioning, for example, encouraged people to stay indoors, leading to a decline in the use of communal spaces like porches, where neighbors once gathered to interact. Similarly, the rise of garages allowed people to drive directly into their homes, bypassing any interaction with neighbors. Backyard fences created private, isolated spaces, contributing further to the erosion of community life.1
Each of these innovations brought convenience and efficiency but at the cost of social cohesion and relational life. Putnam’s argument serves as a powerful cautionary tale for us as we consider the role of AI in education. What unforeseen costs might come with the convenience and efficiency AI offers? What impact will AI have on the relationships, character formation, and communal aspects central to Christian education?
This chapter explores three specific risks that we must consider as we navigate the integration of AI into education. Much more could and will be said, but I think these risks highlight areas where we need to exercise wisdom and prudence, ensuring our educational practices remain faithful to the deeper purposes of formation and human flourishing.
Risk #1: The Idolization of Efficiency
Efficiency is often seen as an unquestioned good, a value to be pursued at all costs. AI offers a range of tools designed to make education more efficient—whether through personalized
1 Robert D. Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon & Schuster, 2000), 63–65, 204–205.
Shaping Minds, Forming Souls 15 learning algorithms, automated grading, or data-driven decision-making. While these innovations certainly improve certain aspects of education, we must be wary of reducing education to efficient processes.
Historically, Christian education has been understood as more than just the efficient transfer of knowledge; rather, it has been seen as a process of formation—a slow, relational, and transformative journey aimed at shaping the whole person. In Christian education, formation has deep roots in the works of early and medieval theologians like St. Augustine of Hippo and St. Thomas Aquinas.
For Augustine, education was not merely about the accumulation of facts or intellectual development, but about the reordering of our loves toward God. In his work Confessions, Augustine speaks of the human heart being restless until it rests in God.2 Education, then, becomes a means of guiding the soul back toward its proper end, namely God Himself. Augustine viewed education as a tool for spiritual growth, shaping not only the mind but also the heart and soul in alignment with God’s will, which he discusses in his work De Doctrina Christiana. For Augustine, true wisdom comes from loving rightly, and the educational process is meant to train us to love God above all else and our neighbors as ourselves.3
Similarly, Thomas Aquinas emphasized the integration of intellectual and moral formation in education. In his monumental work, Summa Theologica, Aquinas argues that the purpose of education is to cultivate both the intellect and
2 Augustine, Confessions, trans. Henry Chadwick (Oxford: Oxford University Press, 1991), 3.
3 Augustine, On Christian Doctrine, trans. Edmund Hill (New York City: New City Press, 1995), 22–31.
16 | Jonathan A. Powers
the will toward the ultimate good—union with God.4 Aquinas believed that education was essential for developing the virtues, especially prudence (right judgment) and wisdom, which would enable individuals to live according to God’s truth. For Aquinas, education was not merely about gaining knowledge but about forming a person’s character to reflect divine wisdom, leading to a virtuous life. He emphasized the importance of contemplation, where the learner moves beyond mere knowledge to a deeper understanding of God’s order in the world.5
A spiritually formative view of Christian education was not confined to the early and medieval church, however. More contemporary theologians have also emphasized the holistic nature of education. John Henry Newman, in his work The Idea of a University, argues that education is about more than specialized training. For Newman, the purpose of education is the cultivation of the whole person, equipping individuals with the intellectual and moral virtues necessary for engaging the world thoughtfully and faithfully. He saw the university as a place where students not only gain knowledge but also learn to live rightly in relationship with others and with God.6
In our own time, theologians like James K.A. Smith have built upon classical ideas of Christian education. Smith argues in his book Desiring the Kingdom that education is fundamentally a matter of shaping our desires and loves.7 He draws on Augustine’s concept of ordo amoris (the order of
4 Thomas Aquinas, Summa Theologiae, I, q.117, a.1; III, q.42, a.4, trans. Fathers of the English Dominican Province (Chicago: Christian Classics, 1981).
5 Ibid.
6 John Henry Newman, The Idea of a University, ed. Frank M. Turner (New Haven: Yale University Press, 1996), 77–79, 100–104.
7 James K. A. Smith, Desiring the Kingdom: Worship, Worldview, and Cultural Formation (Grand Rapids: Baker Academic, 2009).
Shaping Minds, Forming Souls 17 loves), suggesting that all education is formative and liturgical in nature, shaping not just what we think but what we love and desire. For Smith, Christian education must aim at forming students’ affections and desires in such a way that they grow in love for God and neighbor. This process cannot be rushed or reduced to mere intellectual exercises—it involves practices, relationships, and the shaping of the heart over time.8
Christian education, then, is a holistic endeavor. It seeks not only to impart information or prepare students for careers, but also to transform the whole person—mind, body, and soul—toward a life lived in alignment with God’s truth. The formation process is inherently slow, requiring time, patience, and relational engagement. As educators, we are called to cultivate wisdom, virtue, and character in our students, helping them grow not only in knowledge but also in spiritual maturity. These thinkers remind us that the aim of education is not simply intellectual development but a journey toward wisdom, virtue, and ultimately, deeper communion with God.
When we prioritize efficiency above all else, we risk losing these formative elements. True education is inherently inefficient. It requires time for contemplation, space for failure, and room for deep reflection. Reading a challenging book, wrestling with difficult questions, or engaging in thoughtful discussions requires time and space—none of which can be “optimized” in a traditional sense. These slow, formative practices are essential for developing wisdom and virtue. AI technologies, with their emphasis on speed and optimization, can inadvertently devalue these aspects of learning, encouraging students and teachers to see education as a means to an end
8 See Smith, Desiring the Kingdom, 25–30, 53–55, 208–213.
18 | Jonathan A. Powers rather than as a transformative journey of growth in wisdom and virtue.
Risk #2: The Obsession with Innovation
Our culture is driven by an insatiable desire for innovation. In the realm of education, this manifests as a constant pursuit of the latest technologies, teaching methods, and tools—often with the assumption that newer is always better. AI fits neatly into this cultural obsession with progress, promising to revolutionize education by making it more personalized, more accessible, and more “cutting-edge.” But innovation, when pursued for its own sake, can come at a cost.
Wendell Berry offers a helpful critique of modernity’s drive for progress in his writings, particularly in his novel Jayber Crowe. Berry’s work highlights how the relentless pursuit of progress and productivity has eroded the relational and communal aspects of life. The characters in Jayber Crowe become more concerned with efficiency and economic gain than with the slow, deliberate work of building community and tending to the land. The protagonist, Jayber, reflects on how the oncetight-knit community begins to dissolve as people prioritize individual success and technological advancements over the bonds that once held them together. The land itself, a symbol of deep, enduring relationships, is reduced to something to be exploited rather than lovingly cared for.9
In education, the pursuit of innovation can lead to a superficial experience, where depth is sacrificed for novelty. This obsession with the new can fragment the learning process, breaking it down into disconnected tasks or bite-sized
9 See Wendell Berry, Jayber Crow: A Novel (Washington, D.C.: Counterpoint, 2000), 133–136, 184–188.
Shaping Minds, Forming Souls 19 modules designed for quick consumption rather than for deep engagement with ideas, texts, and relationships.
A classical model of education known as the trivium offers a stark contrast. The trivium consists of three stages: grammar, logic, and rhetoric, each designed to build upon the other. In the grammar stage, students learn the foundational building blocks of knowledge—facts, language, and structure. This is where they accumulate information and understand the basic elements of a subject. In the logic stage, students are taught to think critically, reason through problems, and engage in debate. This is the stage where they begin to analyze and synthesize the information they have gathered, learning how to form coherent arguments. Finally, in the rhetoric stage, students learn to communicate their ideas persuasively, applying their knowledge and reasoning to real-world situations.
The purpose of the trivium is not speed or innovation, but the gradual development of wisdom, character, and intellectual depth. It encourages students to engage deeply with material, cultivate discernment, and grow in their ability to think and communicate clearly. This classical model contrasts sharply with a heavy emphasis on efficiency. The trivium’s slow, deliberate process fosters the kind of intellectual and moral formation that can get lost in an obsession with innovation and progress. Any pursuit of wisdom and character development requires time, patience, and a focus on the things that endure.
Risk #3: Fragmentation of Community
Another significant risk AI poses to education is the potential fragmentation of community. From a Christian perspective, education is not merely about imparting knowledge; it is a communal endeavor aimed at the holistic formation of
individuals. Learning in Christian education is deeply relational, achieved through mentorship, meaningful dialogue, shared experiences, and mutual accountability. These aspects are essential for spiritual growth and moral development, reflecting the biblical understanding that we are created for community.
The concept of community is central to Christian theology, grounded in the relational nature of the triune God. God exists in perfect communion: Father, Son, and Holy Spirit. Humanity, made in the image of God, is designed to live in relationship with God and with one another. This idea extends directly into Christian education. As John Wesley emphasized, there is no holiness but social holiness.10 Wesley understood that we cannot be fully formed in isolation; spiritual growth and moral development are deeply communal processes. We learn to love God and neighbor not in isolation but in the context of relationships, where we are shaped by our interactions, our shared struggles, our annoyances with each other, and our mutual encouragement.
AI, however, introduces the risk of shifting education toward a more individualized, isolated experience. As students increasingly interact with machines rather than with teachers or peers, the relational dimensions of learning may be weakened. This isolation undermines the very foundation of Christian formation, which is built on relationships with God, with one another, and with creation. A community of learners fosters growth in a way that cannot be replicated by technology, no matter how advanced. The danger is that AI can turn education into a transactional experience, where information is transferred efficiently, but the deeper, relational work of formation is lost.
10 John Wesley, Hymns and Sacred Poems (repr., London: William Strahan, 1739), viii-ix.
Wesley’s idea of social holiness highlights the need for community in the pursuit of holiness. We are not meant to journey toward Christlikeness alone. In the same way, education—especially Christian education—should resist the pull toward isolation and instead emphasize community as the context for growth. Learning in community allows students to experience the support, correction, and encouragement of others, which is crucial for both intellectual and spiritual formation. The presence of a mentor, teacher, or peer provides accountability, models wisdom, and nurtures the kind of growth and humility that cannot happen in isolation.
Christian education, therefore, has the responsibility to resist the tendency toward individualization that AI can promote. While AI can enhance some aspects of education, such as providing personalized learning paths or automating administrative tasks, it cannot replace the profound impact of human connection and the shared journey of learning. Our classrooms must continue to be spaces where students engage deeply with their teachers and peers, where relationships are prioritized over efficiency, and where the slow, transformative work of personal and communal formation can take place.
In a world increasingly shaped by technology, Christian educators must safeguard the essential role of community. As Wesley’s concept of social holiness reminds us, we are not just individuals pursuing our own goals; we are members of a body, growing together into the fullness of Christ. In this way, education becomes not just a means to knowledge, but a shared journey toward wisdom, virtue, and spiritual maturity.
Conclusion: A Call to Thoughtful Engagement
As we stand on the edge of an AI-driven future, we cannot afford to embrace new technologies without thought. The stakes are too high. Just as Robert Putnam warned of how innovations, adopted without foresight, lead to the erosion of community life, we too must be vigilant. AI has the power to reshape education in profound ways, and if we are not intentional, it could fracture the relational and formational heart of our mission.
Our task as Christian educators, therefore, is to think carefully about the ethical and theological implications of new technologies. We cannot let efficiency and innovation blind us to the deeper work of shaping souls for Christ. Therefore, we must ensure that our use of AI aligns with a robust theological anthropology, one that sees humans as beings created in the image of God, capable of growth in wisdom, virtue, and community. We must remember that education is not only about preparing students for jobs or degrees, but also about shaping their hearts and minds to reflect the character of Christ.
As Christian educators, let us commit together to thoughtfully consider how we can harness the possibilities of AI without losing the essence of what it means to teach, mentor, and shape lives. Together, we must uphold the values that form the core of our calling, ensuring that technology serves our mission rather than distorts it.
Chapter 2
Recovering the Language of Connection in an Age of Connectivity
Daniel H. Strait, Asbury University
My one, perhaps only, qualification for even speaking at this Colloquium is that I am a student of language, not just of words, but of the entire range of human expressiveness. Reading literary texts includes making sense of enacted meanings, speech acts, narrative patterning, and poetic form. One cannot read the poetry of George Herbert, as I do on a regular basis, without an attunement to gestures, embodied practices, liturgies, and rituals. Maurice Merleau-Ponty reminds us that all language is gestural.1 Much of want I want to say in this essay has to do with language and our need to re-discover a language of connection in an age of “connectivity” and machine learning.
I want to address why we are here, first by way of a few remarks about technology in general; second, by offering a placing shot about that form of technology we are calling generative AI, in the specific form of artificial intelligence chatbots, such as ChatGPT and its iterations; and, finally, I will close, not by proposing solutions, but rather by describing a convergence of three things in my own thought life that form (for me, at least) a view of language that exposes the synthetic view of language ChatGPT trades on—and the implications of that trade. This convergence consists of my reading of two Romantic poets and the Catholic philosopher Charles Taylor and organizes around a view of language that is creative and constitutive, vitalistic and alive to the inexhaustible meaning
1 See Merleau-Ponty's Phenomenology of Perception (Routledge, 1962). This idea permeates Merleau-Ponty's view of language.
24 | Daniel H. Strait of the world. ChatGPT creates the illusion that its technology makes everything exhaustibly accessible. It does so without moral limits, or without the fidelity of words to the practical culture within which language is one resource and discipline. “Language’s nature,” says Wendell Berry, “is to turn outward to the world, to strike its worldly objects cleanly and cease to echo— to achieve a kind of rest and silence in them.”2 My concern, which is also my claim, is that ChatGPT is so easy to use and “efficient” that it will have an increasingly instrumentalizing effect on language. The irony is that the more we habituate to ChatGPT, and the more we vacant language, the more we risk cutting ourselves off from the discoverable orders of the world that language should faithfully serve.
I want to begin with a general comment about technology. We are not here to decide whether we are “for” or “against” technology. Technology is as old as civilization. It is also inextricably bound up in our daily lives. If one rakes leaves in the fall or throws a football, if one drives a car or travels by train, orders a pizza online or fishes with fly rod in a Montana river, he or she depends on some form of technology. Our university campus relies heavily on sound stages, projectors, labs, screens, works stations, Wi-Fi, furniture, not to mention on the virtual places we create using Zoom, Microsoft Teams, and Google Hangouts. Technology assists in the work of the community. Missing, though, is more conversation about how the tools of our work cultivate productive communal and cultural practices. Our Amish friends to the south, in Crab Orchard, Kentucky, depend on technology too. They just use it sensibly and humanely in relation to the moral needs of their community. They live deliberately, rooted in the land, and connected, communally, to each other through work, which pays a debt to God. They reject
2 See Wendell Berry, Home Economics, (North Point Press, 1987): 79-80.
Recovering the Language of Connection 25 the attitude that technological progress, in whatever form, is not only inevitable, but also, in fact, “progress” and therefore good.
My second point, more specifically, is that generative AI in the form of ChatGPT furthers the myth of technological progress and heightens our persistent anxiety that we are falling behind and have no time, except technologically configured time. ChatGPT appeals to us because, among other things, it promises to save us time, but its “efficiency,” a badge of honor for AI text generators, works against asking more deliberate questions about how we use generative AI technology and for what purposes. Asking such questions might take time, even prove uncomfortable, but it would exert a productive pressure on us to examine our motivations and practices for using AI technology. In questioning our use of ChatGPT, for instance, we might discover our complicity in lethal economic practices, consumerist exploitation, corporatization, and environmental negligence. Using ChatGPT is not a neutral activity. “A new machine is a new idea,” writes Lionel Basney, which means it is both “objective and hermeneutic.”3 Machines shape our understanding of the world and the language we find to connect to the world. “Once we had made a clock,” adds Basney, “we began to imagine the whole universe as a clock.”4 Are we reimagining the universe as a cosmic chatbot?
Our discussion of ChatGPT needs to acknowledge that our accelerated technological progress fuels our consumptive global economy. We might even need to ask why this does not seem to bother us much. Part of the problem is that we are numbed by the inevitability of “progress,” coupled with our
3 Lionel Basney, “The Schoolroom in the Factory: Reflections on Technology and Its Place in Seminary Education,” Calvin Theological Journal 34 (1999): 421.
4 See Basney, 421.
belief--irresponsible, I think--that there is nothing we can do about it. One hears Dostoevsky’s Underground Man: “just you try to object.” The conclusions of Silicon Valley entrepreneurs, financiers, and techno-utopian enthusiasts insist that AI technology is our destiny. It promises, as it develops, to make our lives better, more productive, more efficient—certainly more profitable. Who, exactly, is profiting from rapidly iterative technological advancements--and how? This question risks being nullified by our hypnotic assent to the idea that questioning technological progress is out of our reach.
Generative AI-ChatGPT lures us with its speed, efficiency, and claims to productivity. This explains why AI technology finds application in a variety of industries across an insanely competitive marketplace: healthcare, finance, marketing, manufacturing, agriculture, cyber-security, consumer retail and e-commerce—and, of course, higher education. None of this is an entirely new problem. Artificial intelligence, in the form of “machine learning,” has been around at least since the Cold War, as Lauren Goodlad and Samuel Baker point out, in their essay, “Now the Humanities Can Disrupt AI.”5 While “machine learning” is not a new problem, the problem for us now is that AI technology is so much closer to us, and more personal, as close as our front or back pockets, and dizzyingly more powerful than we ever imagined.
ChatGPT poses substantial challenges, then, but it also creates opportunities. Many of us, certainly our students, will work, by choice and necessity, with AI-ChatGPT in some form or another. My conviction is that the people who learn outside the system of AI will work best and most responsibly with it. To
5 Lauren M. Goodlad and Samuel Baker, How the Humanities Can Disrupt “AI” https://www.publicbooks.org/now-the-humanities-can-disruptai/.
Recovering the Language of Connection 27 prepare ourselves, and our students, to work with it, we need to think critically about language. One large province of language use is practical and technical. We depend on the instructional value and function of designative language—words and phrases that work for specific ends. Manufacturing processes, software development, pharmaceutical innovation, scientific discovery, public school and university administration, safety protocols, and stop signs, all depend on using language for practical purposes. The legal profession has a technical and specialized language all its own, as do professions in allied health, medicine, and engineering. Beyond its designative function, however, language is also constitutive, as Charles Taylor has pointed out.6 My main claim is that, in this age of connectivity, we need to recover a more constitutive view of language, that is, a language that shapes us and connects us with a God-ordained reality, with others, even with ourselves.
My final thoughts, then, have to do with a convergence of the ideas of three writers, two Romantic poets (Wordsworth and Blake) and a Catholic philosopher, Charles Taylor. This convergence has returned me to an organic, and vitalistic, view of language that generative AI fails to account for. I experienced the first part of this convergence (for the first time) many years ago, in a British Literature II course, while reading the poetry of Wordsworth, whose poetic language makes us feel life close up, “Felt in the blood, and felt along the heart,” as Wordsworth writes in “Tintern Abbey.” The poem unfolds through the interrelationship between memory, morality, and the mystical connection with life—all part of probing the pleasures of the
6 See: The Full Shape of Human Linguistic Capacity (Cambridge: Harvard University Press, 2016); See also Cosmic Connection: Poetry in an Age of Disenchantment (Cambridge: Harvard University Press, 2024).
natural world. Yet one senses in the poem a reality beyond the temporal life:
Nor less, I trust,
To them I may have owed another gift, Of aspect more sublime; that blessed mood, In which the burthen of the mystery, In which the heavy and the weary weight Of all this unintelligible world, Is lightened:—that serene and blessed mood, In which the affections gently lead us on,7
I recall quite vividly the feeling, the affective bonding, I experienced when I read these lines as a 19-year-old. Returning to them now, not only do I recognize the wonder of that earlier moment, but in the fusion of then and now, something else emerges: a new awareness, a deeper sense of connection, with time and its unfolding, with place, and with others. The poem not only describes the moment, or experience; but it also enacts it, and so do we when we read it. Through the language of the poem, the mind revives, and the body feels, in a revitalized and revivified relationship with time and place. As the past interacts with the present in an epiphanic moment (“spots of time,” as Wordsworth called them), the poem discovers, even in its struggle--especially in its struggle—the “food for future years”:
And now, with gleams of half-extinguished thought, With many recognitions dim and faint, And somewhat of a sad perplexity, The picture of the mind revives again: While here I stand, not only with the sense Of present pleasure, but with pleasing thoughts
7 William Wordsworth, “Lines,” English Prose and Poetry, ed. Russel Noyes, (New York: Oxford University Press, 1956); p. 259-260.
Recovering the Language of Connection 29
That in this moment there is life and food
For future years.8
For Wordsworth, poetic creation participates in a dynamic relationship with creation. In remarking on a similar point, Charles Taylor, whose work I will comment on in a moment, observes that, for Romantics, “Our creations reveal what is there, and reconnect us with it.” 9
The second part of this convergence happened in the same class. In reading the poetry of William Blake, I encountered for the first time a writer who had something to say about the effects of mechanization on the world—and on the human person, whom he saw as divine. He blamed Locke and Newton for their part in a cosmic travesty of turning human persons into machines. For Blake, Lockean empiricism turns the mind into a camera—into a machine. “We are led to Believe a Lie / When we see not Thro’ the Eye,” wrote Blake, who takes aim at a world securely “Locked” (Blake’s pun, not mine). Blake felt that imagination could restore this damage.10 Anything that threatens to turn a human person into a machine was a nightmare of horrors for Blake.
Blake detested Newton along the same lines, for rendering the world a coldly mechanical place—a dead world of narrowing vision. In a letter to Mr. Thomas Butts, Great Marlborough Street in London, dated November 22, 1802, Blake described Newton as suffering a “Single vision,” one devoid of the power of the vitalizing imagination:
8 See Wordsworth, p. 259-260.
9 Charles Taylor, The Language Animal, (Cambridge: Harvard University Press, 2016): 244.
10 See James C. Evans, Blake, Locke, and the Concept of “Generation.” https://bq.blakearchive.org/9.2.evans.
Now I a fourfold vision see
And a fourfold vision is given to me
Tis fourfold in my supreme delight
And three fold in soft Beulahs night
And twofold Always. May God us keep From Single vision & Newtons sleep*11
Both Blake and Wordsworth attempted to recover a more organic view of the universe. As a result, they sought out a language animated by the power of imaginative visioning. They wanted to experience a new connection with the world, a cosmic one, not an instrumental one, as Taylor points out in his recent book, Cosmic Connections: Poetry in an Age of Disenchantment, a companion text to his earlier book, The Language Animal.
My re-reading of Wordsworth and Blake happened to converge with my reading of both of Taylor’s books. While Taylor’s The Language Animal explores, at length, our daily need for designative and instrumental language, he concludes by arguing for the need to recover a more intimate connection with the world with the help of Romantic poets who rely on a more creative and constitutive language, whose vision carries them beyond a representational theory of art. The Romantics pursued a language full of vitalism and rejected dead language which, as Taylor puts it, “simply designates things that everyone can see, and allows them to manipulate them, totally ignoring their sign character.”12 A reconnection with the world through
11 See John J. Skillman, “Presidential Address: ‘Single Vision and Newton’s Sleep,’” Presidential address: “Single vision and Newton's sleep,” Journal of Vascular Surgery, 25. 6, 969-974. https://www.jvascsurg.org/article/S07415214(97)70119-8/fulltext.
12 See Taylor, The Language Animal, p. 344.
Romantic poetry reminds us of our need for real connection, “cosmic connection,” as Taylor calls it.13
To conclude, while we do not write business memos or legal briefs or strategic plans in poetry, though ChatGPT might gamify us in that direction, Romantic poetry reminds us that language does not seek to “handle” the world, like one more object for our consumption. “The world is too much with us; late and soon, / Getting and spending, we lay waste our powers;—,” wrote Wordsworth.14 Poetry promotes “cosmic connection” through “subtler languages”—"languages of insight,” as Taylor says. Language locates us in what he calls an “interspace,” where we discover the world in its dynamic unfolding, and its felt depths—its “resonance,” as he says, which emerges through the connection and interweaving of the world with the poet’s act of creation.” 15 The more we habitually vacate our language and rely instead on the processed and re-processed language of generative AI, the more we risk losing touch with what connects us so vitally to God, to the world, and to each other.
13 See Taylor, Cosmic Connections: Poetry in an Age of Disenchantment, (Cambridge: Belknap-Harvard, 2024.
14 See Wordsworth, “The Word Is Too Much With Us,” p. 317.
15 See Taylor, Cosmic Connections, p. 51.
Chapter 3
AI and Human Anthropology: Identifying Conflicts
Bill Arnold, Asbury Theological Seminary
I know we are taught not to begin a class or presentation with an apology. But I am going to make an exception here. I nearly declined Dr. Brown’s invitation to speak today because I have so much to learn about this topic, and hardly think I have much to contribute. Almost simultaneous with the invitation was a “security tip” sent to the Seminary community on August 20 by our IT folks with this definition (independently of the efforts of this colloquium): “Artificial Intelligence (AI) is a technology in which machines are able to simulate human intelligence. Unlike humans, AI has the ability to process large amounts of information very quickly.” I remember thinking, “Wow, the Orwellian Future has arrived.”
Slowly I began to realize that almost all of us can claim we know too little to speak at this colloquium, at least as AI relates to the important topic of identifying conflicts for us in higher education and more particularly for higher education in the Christian/Wesleyan context. So I am going to offer a few very brief remarks, and lean heavily on others for the implications of these observations.
With regard to Christian anthropology, we must all begin with a question about what all this means for our understanding of the imago Dei (the image of God), which is such an important statement in the Bible’s understanding of humankind (Gen 1:2627; 5:1-3; 9:6). Curiously, however, that theme of the “image of God” goes mostly undeveloped and underdeveloped in the
34 | Bill Arnold Bible itself.1 Yet the rest of Bible does have lots to say, of course, about humans and our broken nature, our salvageable--yes, redeemable lives through the love and grace of God, and ultimate salvation from our brokenness by means of the life, death, and resurrection of Jesus. Nevertheless, the topic of the imago Dei has generated so much speculation and philosophizing among Christian theologians for two millennia that I can hardly hope to say much about it that would be instructive for us in ten minutes. Simply put, the specific details of the image are lost on us, or better, they are so rich in theological understanding as to be difficult to conceptualize.
I leave aside this important question for now to attempt to bite off something a little more chewable and perhaps helpful in our discussions. I offer here simply an observation on social theories of personhood in the pre-Hellenistic, ancient world of the Old Testament. First, we need to acknowledge that it is widely and commonly assumed in Western modernity that every human individual possesses both a private and a public self.2 The private or inner self—it is assumed in this worldview—is the one that is true and authentic, whereas the public or social self is a mask that hides the real person. The public sphere is thought to put restraints on individual freedoms, shackling the private self because society imposes limitations on one’s freedom, whereas the private sphere alone allows for full expression of the individual. One’s public self is an actor that
1 Brent A. Strawn, "From Imago to Imagines: The Image(s) of God in Genesis," in The Cambridge Companion to Genesis, ed. Bill T. Arnold, Cambridge Companions to Religion (Cambridge: Cambridge University Press, 2022), 211-35, esp. 211-12.
2 For some of what follows, see Karel van der Toorn, "Public Self and Private Self in the World of the Hebrew Bible," in With the Loyal You Show Yourself Loyal: Essays on Relationships in the Hebrew Bible in Honor of Saul M. Olyan, eds. T. M. Lemos, Jordan D. Rosenblum, Karen B. Stern and Debra Scoggins Ballentine, AIL 42 (Atlanta: SBL Press, 2021), 99-112.
AI and Human Anthropology 35 hides the inner, authentic self. It is not uncommon today to hear someone express the idea that the roles we play in public are foreign and artificial to us because we are following a script written by others, whereas we are only true to ourselves when we are alone or with those closest to us. And this despite the fact that the truth about one’s self is perhaps the most difficult and indeterminate thing for a person to know with certainty. As Harry G. Frankfurt contends,3 By contrast, the sociohistorical context from which the Bible emerged did not share this conviction that the private self was the real self, or even that there was an essential dichotomy between private and public selves. Karel van der Toorn asserts that “The modern cult of authenticity was foreign to their frame of mind?4
Instead, in the Old Testament, it is the public self that is real, or better; The Israelites seemed to have made no distinction between the private and the public self. Neither one nor the other was authentic or inauthentic but were two aspects of the human person. And yet the side of the person that mattered was public, so that one’s identity (and thus fulfillment and real meaning in life) was located in the group one belonged to (family, clan, people). This ancient conception of the human person was central in the creation of “the people of YHWH/ God” ('am YHWH; Numbers 11:29; Numbers 16:41 [Heb 17:6]; 'am elōhîm; 2 Samuel 14:13). The relevance of this reality about ancient Israel is that (1) it anticipated and informed the way the first Christians thought of the Church (ekklēsia), as a corporate reality that transcends ethnicity, cultural status, or any other marker of personhood, and (2) it offers to us today a model for thinking about formational education (making a distinction
3 Harry G. Frankfurt, On Bullshit (Princeton, NJ: Princeton University Press, 2005), 66-67.
4 Toorn, "Public Self and Private Self," 111.
| Bill Arnold here with informational education). I suggest our already unconnected and extremely individualistic culture works against the notion of the Church. And, the question we must honestly and boldly address today is whether the formational education we are called to implement and that we dream is still possible is impacted by the notion that a “machine simulates human intelligence” to rephrase the Seminary’s email from August. What is the Asbury University or Asbury Theological Seminary graduate capable of doing on their own, if their education is shaped in some way by AI? Is that tool really only aiming at an informational education—if that—rather than a genuinely formational education? How much longer will we be able to convince ourselves that what we are doing is really formational? What is the self-standing Christian or pastor capable of thinking and doing, if they have been shaped in their education by AI? How do we respond in a way that is positive but proactive in this liminal moment of IT history?
Chapter 4
Clothing Ourselves with Virtue While Outsourcing Communication: A Challenge
Claire Brown Peterson, Asbury University
In the early days of the 2024 Paris Olympics, American audiences encountered and quickly lambasted an ad for Gemini, Google’s AI chatbot. In the ad, a father asks Gemini to write a fan letter from his young daughter to the hurdler Sydney McLaughlin-Levrone. The public outcry was vocal enough that the ad was quickly pulled.1 It seems that the American public took offence at the underlying message—that the actual fan letter of an actual child was not good enough, but Gemini was there to help. The problems involved in the “Dear Sydney” flop may seem isolated, but I suggest that they are connected to a larger challenge for the development and exercise of virtue that occurs when generative AI tools are habitually used in the context of personal communications. I will first provide a quick overview on the nature of the virtues and the way they are acquired and then explore how these features of the virtues are relevant to certain intended applications of generative AI.
Virtue and its Development
The simplest definition of a virtue is that it is a morally good character trait. Honesty, kindness, courage, and temperance readily come to mind as examples, but these four are by no means exhaustive of the list of virtues and, indeed, we need not assume that there is some perfect master list to be found
1 Todd Spangler, “Google Pulls Controversial ‘Dear Sydney’ Gemini AI Ad From NBC Olympics Rotation After Backlash: ‘Who Wants and AI Written Fan Letter??’,” Variety, August 2nd, 2024, https://variety.com/2024/ digital/news/google-controversial-dear-sydney-gemini-ai-ad-pulled-nbcolympics-1236094028/.
| Claire Brown Peterson identifying all human virtues. The most obvious feature of the virtues is that, as morally good character traits, together, they provide dispositions to right action. People with the virtues tend to reliably do the right thing. Crucially, however, as character traits, virtues are “whole person” orientations influencing what we do, why we do it (i.e., motive), the things we tend to notice, where and to whom we direct our attention, and the feelings that arise in certain situations as if by second nature. The individual virtues thus turn out to be quite textured. 2 Specific virtues are also often more complex than initially meets the eye. Consider the virtue of gentleness. When we initially think of “gentleness,” we might imagine a Mister Rogers-type figure who is soft-spoken and rarely if ever angry. While Fred Rogers—the actual person—may well have been an exemplar of gentleness, trouble comes when the Mister Rogers we are imagining turns out to be little more than a caricature.
The “whole person” orientation aspect of the virtues can make their possession appear a matter outside the human will. The idea that humans can only control their actions and not their emotions—with the upshot that one’s feelings are not subject to moral judgment—has achieved something close to tenet status for many in American society. By contrast, the virtue ethics tradition holds that particular emotional responses can be judged as morally bad or good (e.g., joy in a perceived rival’s failure is typically a moral failing). While it is true that our emotional responses in particular situations are outside of our immediate volitional control (this is the truth behind the “you can’t control how you feel” mantra), this fact does not render emotions completely disconnected from the
2 For an accessible introduction to virtue ethics, see Rosalind Hursthouse, On Virtue Ethics (Oxford University Press, 1999), especially pp. 108 -120.
Clothing Ourselves with Virtue 39 human will. Indeed, the acquisition of virtue, including virtue’s emotional dimensions, is the result of both the Holy Spirit and repeated human actions and inactions. We can point to the Biblical foundations of this human role in reorienting our hearts in Colossians, “Put to death, therefore, whatever in you is earthly: sexual immorality, impurity, passion, evil desire, and greed (which is idolatry)…clothe yourselves with compassion, kindness, humility, meekness, and patience…above all, clothe yourselves with love, which binds everything together in perfect harmony.”3
While there is no single way that this human aspect of “clothing ourselves” with the virtues plays out, the short version of the “how” of virtue development is that we habitually and repeatedly make certain choices. Eventually, those repeated choices change the parts of our character that are not under our direct intentional control—our desires, our impulses, what we notice, what we naturally pay attention to. In the case of the virtue of love, for example, we clothe ourselves with a love we do not yet have by trying to do the loving thing even when we do not feel it—by, among other things, taking the time to really listen to and be present with people, attending to what is weighing them down, and finding ways we can help alleviate some of their burdens. Eventually, thanks to the work of the Holy Spirit, this becomes more like second nature to us.
The process of virtue-development is importantly a communal one: Paul rather boldly says, “be imitators of me, as I am of Christ.”4 Ideally, when we are not sure what to do, we turn not to our caricature of Mister Rogers or what immediately comes to mind when we think of gentleness or courage but to the people we know to be gentle and courageous. We identify 3 Colossians 3: 5, 12b, and 14 (NRSVUE).
4 1 Corinthians 11:1.
| Claire Brown Peterson our Christ-like exemplars and we try to imitate them in their imitation of Christ. In actual practice, this process is quite messy. We can fail miserably in our attempts at imitation, and real people can get hurt as a result.
How might the use of AI tools present challenges for the development of virtue?
The underlying pitch of generative AI is that of quick, easy, excellence: decrease your own personal involvement because AI can do a better job at the thing you are trying to do, and it can do it more quickly. The “Dear Sydney” ad previously described exemplifies this pitch, but the most unique aspect of the ad when it comes to this pitch is that a child is the one whose words are being pushed aside.5 This deferment of personal investment in the name of quality and efficiency raises concerns when it comes to some of the likely effects of AI tools on virtue development, particularly in the domain of personal human communication.
Jesus told us that “it is out of the abundance of the heart that the mouth speaks.”6 But what if what comes out of my mouth (or text message) purports to express my thoughts but
5 This is no accident. In identifying the goal of AI, the most widely used AI textbook identifies two main ways of conceiving of the type of intelligence that is trying to be achieved: the intelligence of a human and ideal rationality. Stuart Russell and Peter Norvig, Artificial Intelligence: A Modern Approach, 3rd ed. (Prentice Hall, 2010), pp. 1-5. If the goal is merely human intelligence, one wonders why an actual human would ever bother with it, and the ready answer appears to be “because I lack the level of human intelligence that this AI program manifests when it comes to this task” and/or “because I do not want to invest the time/ effort I would need to make use of my own human intelligence.” If the goal is ideal rationality, then the benefit of turning to AI is obvious: AI can do this better than I could even if I invested great time and effort. No matter what kind of intelligence AI promises, then, the most natural impetus for its use will be efficiency, quality, or some combination.
6 Luke 6:45.
Clothing Ourselves with Virtue 41 was largely written by a bot—is that something that “overflows from my heart”? Obviously, there are multiple levels and dimensions of agency that can be involved or suspended when we use AI to help us think through a position or express our ideas. Given this range even within the limited domain of personal communication, assessments of such uses quickly get complicated.
Even so, when generative AI can quickly give us suggestions and words that sound impressive and maybe even are perfectly good things to say—they are the sort of thing a kind, compassionate, thoughtful person would say—we may be bypassing the messy process of trying to clothe ourselves with virtue and instead habituating ourselves to largely checking out. Consider, as a recent Atlantic article did, the personal attention involved in sending an AI-generated message of support to a friend experiencing divorce.7 Lest we think such events are too infrequent to be of concern, we need only realize that every instance of written communication that is emotionally difficult to write is ripe for such use: the text asking for donations for one’s child’s basketball team, the email thanking VBS volunteers, and the letter of apology to a work colleague. To be fair, this problem of going through the motions is not a new one: the process of virtue formation has always involved the temptation to substitute the semblance of virtue for the development of the actual thing, and this has always been a tricky balance. Generative AI may be seen as multiplying a temptation that has always been present.
7 Matthew Schnipper, “My Deepest Condolences. Signed, Chat GPT,” The Atlantic. October 3rd, 2024. Interestingly, while the recipient of the text in question felt betrayed, the author of the piece was less worried, his first thought being simply, “At least the person texted back.”
Perhaps, though, we are careful to use generative AI not as a crutch and rather as a learning resource. (So, I am not just letting Gemini think for me; I am trying to learn from it. I am actively trying to change). This strikes me as in-principle and in many domains a perfectly reasonable thing to do. What needs to be acknowledged, however, is that when we use AI to substantially shape how we think through and talk about difficult ethical and interpersonal issues, we are essentially imitating ChatGPT as it imitates its training data. That is not necessarily a bad thing. But at the very least, it behooves us to realize that that is what we are doing. I suspect we will very easily forget this fact.
Speaking is of course only one side of communication; there’s also listening. Consider the experience of feeling dismissed because the person with whom we are talking seems to be largely ignoring our actual words. Five seconds into the conversation, our “listener” has decided what we are trying to say and is already crafting a response. Most of us have been on both ends of this experience. Absent special circumstances, when a human being who wants to be heard is speaking, we cannot routinely outsource our attention to a bot while at the same time exercising love for that human being and more generally cultivating the virtue of love. How might we be tempted to do just that? “Hey, AI assistant, please summarize this article for me.” “Please summarize this email for me.” “What’s the gist of these MarcoPolos?” Our interactions with our fellow-humans— including people we call friends—are increasingly conducted at a distance and asynchronously. In that context it will be very easy to habitually ask for summaries and responses and never stop to wonder if we have actually loved our neighbor by giving them the attention they crave.
Chapter 5
The Self as Knower in an AI World
Paul Nessleroade, Asbury University
When navigating the challenges that come with the proliferation of AI tools, we should be encouraged to first step back and look at ourselves, the agent of navigation. Far too quickly, it can be argued, we place all our attention on understanding the functionality of new AI tools, presuming we already know enough about ourselves. But a more soberminded assessment of who we are as knowers will allow us to see issues related to AI more clearly.
Let us start with the question of how a self comes to know itself; what motivations govern the regulation and management of our sense of self. How do we seek out and gather up information and build a picture of who we are? What are our wants and needs? Theorists suggest there are at least three motives, and that when more than one is engaged, there is a predictable pecking order.1
First, we have a need for self-assessment, a need to find accurate self-relevant information; we want to know what is true about ourselves.2 For instance, when given an opportunity
1 Dauenbeimer, D. G., Stablberg, D., Spreemann, S., & Sedikides, C. (2002). Self-enhancement, self-verification, or self-assessment: the intricate role of trait modifiability in the self-evaluation process. Revue internationale de psychologie sociale, 15(3-4), 89-112. Cf: Sedikides, C.; Strube, M. J. (1995), The Multiply Motivated Self. Personality and Social Psychology Bulletin, 21(12): 1330–1335.
2 Sedikides, C. (1993). Assessment, enhancement, and verification determinants of the self-evaluation process. Journal of Personality and Social Psychology, 65(2), 327–338. Cf: Sedikides, C., & Strube, M. (1997). Self-evaluation: To thine own self be good, to thine own self se sure, to thine own self be true, and to thine own self be better. Advances in Experimental Social Psychology, 29(29), 209–269.
| Paul Nessleroade to choose between two different measurement tools designed to assess some personal trait or disposition, if we are told that one tool is more accurate than another, we tend to opt for the more accurate one. We value accuracy. So far, so good.
But there can be at least two other motives aroused within a given situation, both of which appear to be more powerful than our need for accuracy. One is the need for verification, for our world to parrot back to us descriptions of ourselves that we already hold to be true.3 When we are presented with information that fits well with our picture of ourselves, that is pleasing. We tend to accept this information with little scrutiny. Correspondingly, it is displeasing and unsettling when we are presented with information that seems contrary to that which we already “know” to be true of ourselves. We tend to criticize this self-relevant information, minimize it, find ways to marginalize it, and ultimately try to discard it as mistaken. Our desire for verification seems to be more powerful than our desire for accuracy. It suggests we would rather be right than be enlightened.
We also have a motive for self-enhancement, a desire to see ourselves in the best possible light within the constraints of a given situation.4 It is activated whenever we are pondering an
3 Swann, W. B., Jr., Pelham, B. W., & Krull, D. S. (1989). Agreeable fancy or disagreeable truth? Reconciling self-enhancement and self-verification. Journal of Personality and Social Psychology, 57(5), 782–791. Cf: Swann, W. B., Jr., & Buhrmester, M. D. (2012). Self-verification: The search for coherence. In M. R. Leary & J. P. Tangney (Eds.), Handbook of self and identity (2nd ed., pp. 405–424). The Guilford Press; Talaifar, S., & Swann, W. B. (2017); Self-verification theory. Encyclopedia of Personality and Individual Differences. Cham: Springer International Publishing. pp. 1–9.
4 Beauregard, K. S., & Dunning, D. (1998). Turning up the contrast: Selfenhancement motives prompt egocentric contrast effects in social judgments, Journal of Personality and Social Psychology, 74 (3); 606–621. Cf: Brown, J. D., & Gallagher, F. M. (1992), Coming to terms with failure: Private self-enhancement and public self-effacement, Journal of Experimental Social Psychology, 28,
The Self as
aspect of the self that is perceived to have value. For example, it is valuable to be a good friend, to be a fair person, and to be effective at what one is doing. Many lines of evidence support the power of this motive. For instance, we engage in an asymmetrical way of interpreting our causal role in successful versus unsuccessful endeavors, readily accepting credit for success (even if we exude modesty in public) but denying responsibility for failure.5 We also tend to be unrealistically optimistic about the future, believing that misfortune is more likely to befall others.6 Freud liked to tell the story of a middleaged couple sitting at the breakfast table, the husband leaning out from behind the morning, paper saying, “you know dear, if one of us were to die early, I think I’d move to Paris.”7 Another line of evidence has to do with the favorable way we interpret our best and worst qualities. In brief, we see our failings as common, but our virtues as rare.8 We also tend to make favorable self-other comparisons. On any generally valued trait, we tend 3–22; Klein, W. M. (1997), Objective standards are not enough: Affective, selfevaluative and behavioural responses to social comparison information, Journal of Personality and Social Psychology, 72 (4), 763–774.
5 Campbell, W. K., & Sedikides, C. (1999). Self-threat magnifies the selfserving bias: A meta-analytic integration. Review of General Psychology, 3(1), 23–43.
6 Fischer, R., & Chalmers-Brown, A. (2008). Is optimism universal? A meta-analytical investigation of optimism levels across 22 nations. Personality and Individual Differences, 45, 378-382. Cf: Waters, E. A., Klein, W. M. P., Moser, R. P., Yu, M., Waldron, W. R., McNeel, T. S., & Freedman, A. N. (2011). Correlates of unrealistic risk belief in a nationally representative sample. Journal of Behavioral Medicine, 34, 225-235; Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39(5), 806–820; Massey, C., Simmons, J. P., & Armor, D. A. (2011). Hope over experience: Desirability and the persistence of optimism. Psychological Science, 22, 274-281.
7 Freud, S. (1899/1997). The interpretation of dreams (A. A. Brill, Trans.). Wordsworth Editions.
8 Krueger, J., & Clement, R. W. (1994). The truly false consensus effect: An ineradicable and egocentric bias in social perception. Journal of Personality and Social Psychology, 67(4), 596–610. Cf: Goethals, G. R., Messick, D.M., & Allison, S. T. (1991). The uniqueness bias: studies of constructive social comparison. In
| Paul Nessleroade to think of ourselves as “better than average.”9 Finally, once learning of these self-serving tendencies, most people believe that they avoid them better than others;10 this is the so-called “ultimate” self-serving bias.11 The robust nature of these lines of evidence suggest that we would rather believe we are good than know the truth.
There is another layer to this problem. It is widely held that a good self is one who seeks to know the truth about oneself. Being a person who holds an accurate picture of themselves is the kind of self we all aspire to be. So, from a self-regulation perspective, this is how self-deception comes about. We are insulated from gaining a clear picture of who we are by the very motivations used to create and regulate that picture of who we are.
Now, of course, our biased self-perceptions do not have run-of-the-place, carte blanche. Reality pushes back. Living
J. Suls & T.A. Wills (Eds.) Social comparison: Contemporary theory and research. Hillsdale, NJ: Erlbaum.
9 Messick, D. M., Bloom, S., Boldizar, J. P., & Samuelson, C. D. (1985). Why we are fairer than others. Journal of Experimental Social Psychology, 21(5), 480–500. Cf: Wylie, R. C. (1979). The Self-Concept: Volume 2. Theory and Research on Selected Topics. Lincoln: University of Nebraska Press; Guerin, B. (1994). Attitudes and beliefs as verbal behavior. The Behavior Analyst, 17(1), 155–163; McKenna, F.P. & Myers, L.B. (1997). Illusory self-assessments—Can they be reduced? British Journal of Psychology, 88, 39-51; Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers? Acta Psychologica, 47, 143–148; White, M. J., Cunningham, L. C., & Titchener, K. (2011.) Young drivers' optimism bias for accident risk and driving skill: Accountability and insight experience manipulations. Accident, Analysis, and Prevention, 43(4), 1309-1315; Cross, K. P. (2006). Not Can, but Will College Teaching Be Improved? New Directions for Higher Education, 17, 1 - 15.
10 Pronin, E., Lin, D. Y., & Ross, L. (2002). The Bias Blind Spot: Perceptions of Bias in Self Versus Others. Personality and Social Psychology Bulletin, 28(3), 369-381.
11 Friedrich, J. (1996). On Seeing Oneself as Less Self-Serving Than Others: The Ultimate Self-Serving Bias? Teaching of Psychology, 23(2), 107–109.
in a community with others has a way of keeping in check the more flagrant versions of our biases. But community life is not a full remedy. After all, communities are made up of collections of selves that have their own group-based biases.12 However, there is also an internal antidote; the cultivation of virtues; the intentional work we do on ourselves to obtain personal qualities that we do not naturally possess. One of the classical intellectual virtues was termed “scientia.” Long before this word was associated with an investigatory method or a profession, it represented the cultivated aspiration to know the world accurately; to root out and excise superstitious thinking and false beliefs.
Critically, our self-serving motivations interfere with not only self-perception, but also perception understood more generally. As we reach out to engage the world, our sense of self extends with us and absorbs into itself that which we contemplate.13 The psychological self is not like the biological self – the psychological self extends as our interests and explorations extend. It includes all that we hope to call mine; in addition to our possessions, it includes our affiliations with others, our attitudes and beliefs, and our theories about how the world works.14 Once our ideas are formed, we become invested in them. Yes, we want them to be correct, but we also want them to be verified and to reflect positively upon us. The very feature of us that prompts us to be curious and to reach
12 Taylor, D. M., & Doria, J. R. (1981). Self-serving and group-serving bias in attribution. Journal of Social Psychology, 113 (2), 201–211. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In Austin, W. G.; Worchel, S. (eds.). The Social Psychology of Intergroup Relations. Brooks-Cole. pp. 33–47.
13 Lancaster, S., & Foddy, M. (1988). Self-extensions: A conceptualization. Journal for the Theory of Social Behaviour, 18(1), 77–94.
14 James, W. (1890). The Principles of Psychology. New York: Henry Holt and Company.
| Paul Nessleroade out to come to know the world around us, namely, an active inquisitive self, brings with it into those explorations all our self-regulation mechanisms, including our biases. As we look out to form political opinions, our need to see ourselves as a good person and our past actions as having been moral and appropriate comes along with us. As we form a theory about other people, the need to see ourselves as good by comparison is there. Of course, these self-serving impulses do not determine our beliefs, but they are a part of the recipe mix. The nature of the knower interacts with that which is to be known. Since the postenlightenment, a rival understanding of the self has emerged and become implanted within the Western imagination. From this view, the self is not plagued with motivational flaws that frustrate our search for truth. This self does not lack formation but rather only information.15 From this perspective, our selfserving biases are the product of ignorance, not arrogance. The self is not ill-formed, only ill-informed. Evidence of this view of the self can be seen in the step-by-step diminishment of moral instruction from the Western academy16 and in the migration of the concept of “scientia,” out of an inner quality to be cultivated and into a method merely needing to be followed and a professional label instrumentally to be worn.17 From this view of the self, great optimism was attached to the information age. It led us to think that if only we knew more about each other, we would naturally come to care more for each other. If
15 Trueman, C. (2020). The rise and triumph of the modern self: Cultural amnesia, expressive individualism, and the road to the sexual revolution. Wheaton, IL: Crossway.
16 Reuben, J. A. (1996). The making of the modern university: Intellectual transformation and the marginalization of morality. Chicago: University of Chicago Press.
17 Harrison, P. (2015). The Territories of Science and Religion. Chicago and London: The University of Chicago Press.
only we knew more about the world, we would naturally engage with it more appropriately.
How one understands the nature of the self has many implications for how one approaches and engages with AI search tools. One significant implication relates to how AI tools supply the self with knowledge about simple, factual questions. It stands to reason that those who hold to a classical understanding of the self would be more likely to bring some skepticism to not only the data distilled from an AI process, but also to the one who is interacting with that data. Let’s consider the question, “Was Dietrich Bonhoeffer an advocate for nonviolence?” Our desired outcome can influence issues related to the framing of the question, definitional issues, selection issues, interpretation issues, and so on. Now consider some more typical “factual” questions that we might try to answer with the help of AI, like, “Are masks effective?” or “Is there a gene for alcoholism?” Questions like these are not easy to answer for a whole host of reasons, although a search engine’s distillation coupled with our presuppositions might entice us into thinking so.
Another implication has to do with our more complex and existential questions. For several hundred years now, we have nurtured the idea that we are very capable knowers, and further that the world out there is capable of being known. The ability of databases to hold and compile vast amounts of information and for search engines to gather and compile this information and provide us with answers to our questions will no doubt lend support to this presumption. But, in time, these machines may also unearth an underground fissure that lies between the kinds of information that these tools can provide (reliable or not) and what it is that we really need to know. AI tools will undoubtedly claim to know how to help us run a profitable business, live a
50 | Paul Nessleroade
healthy life, and nurture a happy family. That is until they do not help us in these ways: until we experience bankruptcy, until we get a dreaded medical diagnosis, or our marriage falls apart and our children disown us. As we try to gather up the pieces of our broken lives, we will be confronted with different and more elusive questions; we will recognize that we have a need for a qualitatively different kind of knowledge.
In both situations, those who view the self as needing formation (or shall we say, transformation) as well as information may be in a better position to apprehend truth when it is available, or at least to know better what we do not know. In a world awash with information, there might be a premium placed on those who have cultivated such a taste and respect for truth that it overwhelms the impulse toward self-verification and self-enhancement.
Certain interpretations of an AI future suggest that intelligence will soon be decoupled from consciousness18 and that the human capacity to know will be fully modeled.19 But reductionistic theories applied to humans have not fared particularly well (e.g., Horst, 2007).20 In fact, the failure of these hopelessly ambitious models has ironically only served to reveal a staggering level of complexity and interdependency underlying each of these phenomena. An argument could be made that the proliferation of AI tools, rather than serving to replicate and make redundant our epistemic processes, will only serve to highlight and underscore the profound mystery surrounding the human ability to come to know. The AI world
18 Harari, Y. N. (2016). Homo Deus. Harvill Secker.
19 Fokas, A. S. (2023). Can artificial intelligence reach human thought? PNAS Nexas, 2(12), 1-5.
20 Horst, S. (2007). The demise of reductionism in the philosophy of science. Beyond Reduction: Philosophy of Mind and Post-Reductionist Philosophy of Science, Philosophy of Mind Series. New York: Oxford Academic.
that we are entering, amidst all the benefits and drawbacks, might also contain a prime opportunity for Christian communities and institutions of Christian higher education to point people toward the One who has made us curious and has given us a capacity to know, at least in part; and who has supplied us with the hope that one day we will know in full, just as we are fully known.
Chapter 6
Embodied Image of God: Implications for AI
A. Sue Russell, Asbury Theological Seminary
Introduction
As a biblically trained anthropologist and an anthropologically trained Biblical scholar, I have had to reconcile faith commitments with the paradigms of my chosen discipline. Anthropology seeks to discover how humans and human culture evolved. As a Christian anthropologist, I ask what makes us uniquely human and how does this differentiate us from the rest of creation? In the emerging technology of AI, this same question may help guide our understanding of the capacities of AI and the uniqueness of humans as the embodied Imago Dei. In this paper I contend that the uniqueness of being human, including the development of human culture, is the reflection of the embodied image which cannot be replicated by AI.
To demonstrate this, I briefly present the anthropologists’ view of what it means to be human and how evolutionary theories fall short in explaining the complexity of human experience. I also reflect on how the biblical narrative shows the image of God being expressed in the creativity, relationality, and moral agency of humans which are the foundation of human culture. Then I discuss how our embodiment of the image of God and development of culture differentiates our human capacity from AI.
Origins of Humankind and Culture
For this paper, when I discuss culture, I am referring to the social and ideological aspects of human experience as well
54 | A. Sue Russell as the role of human agency within culture. Biblical scholars and anthropologists study the same phenomenon, humans, and human culture, but from different starting points. Both disciplines ask similar questions: what makes us human, what accounts for the seemingly universals in human behavior and capabilities, and what accounts for the differences between cultures? Both Biblical scholars and anthropologists seek to answer questions about violence, war, injustice, and deviant behavior that seem contrary to the good of humankind. They also ask questions about altruism, kindness, and self-sacrifice that contradict preservation of self. Both biblical scholars and anthropologists ask the question, “What does it mean to be human?” Though both disciplines ask the same questions, the difference is the metanarrative in which evidence is filtered. Though these metanarratives result in different conclusions, anthropology provides concepts to explore what it means to be created in the image of God.
The modern discipline of anthropology was born in the cultural milieu of rapid industrialization, culture change, and the rise of scientific reasoning which included Charles Darwin’s theory of biological evolution. This was also the age of colonization when Europeans were learning about exotic cultures from government workers, missionaries, and explorers.1 Many of the early theories sought to explain the differences of the cultures that were observed and used Darwin’s concepts to construct theories of social evolution. According to these theories, cultures started at a primitive level and evolved into ever increasing complexity with Western “civilized” society viewed as the apex of this evolution.
1 Brian Howell and Jenell Williams Paris, Introducing Cultural Anthropology: A Christian Perspective (Grand Rapids: Baker Academics, 2011), 20.
Embodied Image of God 55
Later theories rejected the idea of unilinear evolution of culture but still had as their guiding question, “How did culture (broadly including social and ideological aspects) evolve in humans?” They attempted to answer the question with theories that sought to understand the relationship between the individual, worldview, and social structure. Through the years, numerous theories arose that more or less emphasized each of these relationships, but none fully answered the question, “From whence does culture arise?”
Theories of evolution fall short in answering this question. Although there is evidence of some characteristics of culture in other animals such as the use of tools and the use of symbols, it is the complexity and meaning-making that render humans distinct. For instance, the complexity of our language, symbol systems, and social institution scannot be explained by evolution. The creation of culture is a uniquely human experience. I propose that the questions of the uniqueness of humankind and the development and complexity of human culture are found in the metanarrative of the Biblical account of creation.
Our uniqueness as humans and the origin of culture arises out of the embodiment of God’s image. Eloise Meneses argues that our uniqueness comes from our special calling as viceregents having stewardship of creation.2 However, the image of God is not only found in our calling but in three important characteristics of the embodied image of God that form the basis of human culture. Each of these characteristics is found in the creation story: our creativity (specifically humankind’s use of symbols); our relationality (our relationship to creation, to God, and to each other); and our moral agency (the ability and 2 Eloise Meneses, Studying the Image: Critical Issues in Anthropology for Christians (Eugene: Cascade Books, 2019), 67.
56 | A. Sue Russell freewill to make decisions with the moral capacity to choose what is aligned with God’s intention for human flourishing).
The image of God in humans is found in the complexity of our symbol systems and our capacity to give them meaning. We create, and in doing so reflect the very nature of God, the creator. “In the beginning God created…And God said, ‘Let there be light,’ and there was light.”3 Each day God spoke, and each day God created. The first task God gave Adam was to name the animals. Adam demonstrated the divine image as he created names (i.e., symbols) for the animals. Humans’ use of symbols is complex, and this complexity cannot be explained by evolution. The use of symbols and rituals to convey complex meaning and emotion is uniquely human. Each time your heart soars to music or your eyes view a masterpiece that takes your breath away reflects the capacity of humans as the embodied image of God to create something that points us back to Him.
The image of God is also found in humankind’s relationality. God is relational, and we are made for relationships. He said, “Let us make man in our own likeness.”4 God placed humankind in relationship with Himself, with each other, and with creation. From these relationships developed kinship systems, political systems, economic systems, and religious systems. Our uniqueness is in the complexity of relationships that form deep human relationships and develop enduring social institutions.
Finally, the image of God is reflected in our moral capacity and agency. Humankind has moral awareness of what aligns with God’s intent for human flourishing and the ability to choose to do it. We also have the capacity to choose to harm
3 Genesis 1:1, 1:3.
4 Genesis 1:26.
Embodied Image of God 57 others and to act in ways that are detrimental to ourselves and others.
Implications for AI
How does understanding the uniqueness of humans help us comprehend AI and differentiate human capacities vis-àvis the capacity of AI? Although AI can perform human-like tasks, AI cannot replicate the embodied image of God. Humans can create; AI can only use what humans have created. AI can learn from vast amounts of data and is able to provide new combinations of this data, but it only compiles what it learns rather than creating entirely new theories. In other words, AI can learn from data, and it can create content from existing data, but it must be trained from existing information which was created by humans.
AI can also imitate human emotions and ways that humans connect, but it does not have the human capacity for relationships. Relationality requires far more than learned data; it is both physical presence and emotional connection. The image of God in us gives us a deep need for belonging and connectiveness. We rely on the relationships formed in creation – our relationship with creation, with each other, and with God-to sustain us physically and emotionally. These relationships give humans a profound sense of purpose and meaning. AI does not need these connections to function, nor is it capable of creating them. And finally, AI does not have an innate understanding of morality nor freewill to choose anything other than what it is programed to do within its certain parameters.
Conclusion
AI is an amazing tool and has the potential to aid humankind in many ways by being able to process data, analyze
58 | A. Sue Russell patterns in data, and make predictions far quicker and more comprehensively than humans. However, we cannot and should not let AI take the place of nurturing our creativity and relationality. We need to nurture the qualities that make us uniquely human, the creativity and relationality that reflect our embodied image of God.
Chapter 7
Identifying Agent and Patient: AI as Moral Subject1
James R. Thobaben, Asbury Theological Seminary
Artificial intelligence (AI) is a technology with stunning potential for improving the human condition and for creating the machinery of human destruction. In actual practice, it may be neither. Definitionally, artificial intelligence is a network of computing machines having the capacity to learn from experience such that it can self-direct in performing increasingly complex tasks. AI functions as a close substitute for humans in performing certain tasks.
Minimally as a tool, AI may be used rightly or wrongly depending on the character and capacity of the humans triggering the use. What, though, if the AI itself becomes something more. To use Kantian language, what if – in some sense – AI moves from being a “means only” to an “end in itself”? In Christian terms, what if AI becomes something more than “mere machine,” an entity capable experiencing good and evil?
To consider this, three categories of AI abuse need to be considered and juxtaposed with the types of AI as defined by capacity. These should be assessed in comparison, so a risk analysis approach is helpful, though not conclusive. Finally, a normative consideration of the moral meaning of “person” must be applied to interpret the risk assessment. First, then, let us consider the capacity of AI.
1 Appreciation for reading and suggesting is given to J. Horstman and M. Thobaben. All errors, of course, remain the author’s.
Types / Levels of AI
AI is not of a singular kind. A list of lesser to greater selfdirection provides some useful distinctions for ethical clarity. (This list, not surprisingly, is similar to, but not identical with, other lists being proffered.) The distinctions are drawn by considering capacity to address events yet to occur:
1. Almost AI: Algorithmic responses that mimic human responses, albeit with far greater speed and precision; the tasks performed are very narrow as is the dialectic with rapid iteration (in other words, it is really only going to provide reductive options depending entirely on assembling and sequencing prior input).
2. Reactive AI: Very weak AI capable of responding to external stimuli in real time; unable to build memory or store information for the future beyond the singular task, but as the machine’s “thinking” is linear. It can process options through decision trees.
3. Narrow AI: AI designed to complete specific tasks / actions; unable to independently learn beyond the narrow field, yet capable of providing options that account for hazard or benefit of outcomes as the machine uses probabilities.
4. General Intelligence / Generative AI: AI designed to learn, think, and cognitively perform actions, including through the use of analogy allowing the crossing of task fields (the grandest form of this is superintelligence AI).
5. Responsive AI: AI that has cognitive capabilities of General Intelligence and can sense human emotions as part of its input, thus becoming “sensitive” in providing
Identifying Agent and Patient 61 appropriate responses; it can use analogy to understand humans (sometimes called “theory of mind AI.”2
6. Self-Aware AI: AI that has cognitive capabilities of General Intelligence and can sense human emotions and has some degree of self-awareness of itself as a distinct entity; this is a machine capable of analogical thinking across categories and has senses inside [machine proprioception] and outside [external sensate gathering].
7. Reproductive AI: AI that has cognitive capabilities of General Intelligence and can sense human emotions and has some degree of self-awareness as a distinct entity and has some degree of “intentional reproductive capacity.”
The last two categories can together be referred to as “Very Strong AI.” Note, though, that this language remains quite unsettled. For instance, most of what is called here “Very Strong AI” might also be called Artificial General Intelligence.
At any of these levels, AI dramatically alters access to and the assembly of knowledge, using algorithms and iterative mechanisms to rapidly sift and process data to provide a preferential selection of options given seemingly countless possible scenarios. Perhaps, at very advanced stages, AI might be able to apply values to novel situations on the basis of such assembled knowledge. In other words, either due to a quantitative increase in capacity or a qualitative change in processing, AI may, in some sense, think and value.
A simple way to describe the process is:
• That data are assembled into facts;
• And facts collated into information; 2 “Types of Artificial Intelligence | IBM,” October 12, 2023. https://www. ibm.com/think/topics/artificial-intelligence-types.
• Then information is assembled in a coordinated manner as subject knowledge; and,
• Knowledge applied with values is conceptualization.
• For Christians, concepts when interpreted with normative values of goodness, beauty, and truth into a singular whole is understood as wisdom.
• Wisdom applied is prudence.
AI is currently capable of data collection, information assembly, and target application of knowledge. Arguably, AI is at the edge of “conceptualizing.” In a sense, AI does that which previously could be done only by the human mind. It cannot, at least not yet, do all that a human does in making a choice, but does do much more than any machine in the past. And what it does do it does much faster than a human could.
This raises not only possibilities, but concerns. The biggest concerns center, not on the processing of bits of facts and components of information, but on the directing of decision-making, perhaps due to unnamed predilection or bias. The overall caution with AI is not simply that raised with “traditional” computing of GI/GO (garbage in / garbage out) as over the last five decades. The danger is not merely one of processing, but of choice selection amplified by the power to control other systems in implementing these decisions.
Certainly, on the basis of all common descriptors used for human cognition, very strong AI would seem to be “thinking” and might be deemed “willful.” Whether it might become “sentient” (or sensing / feeling) is a trickier determination, especially if that “feeling” is understood as going beyond the physical to having a notion of empathy. A Very Strong AI having “sensors” might be nothing more than a reducible feedback
Identifying Agent and Patient 63 system that without consciousness triggers responses, like an exaggerated version of a backup camera linked to a dashboard alarm and braking on a vehicle. Yet, if AI could become capable of “feeling” its own pain or anticipating such in another, then the decision-making would no longer be just a set of automated reactions, but something quite similar to the processes involving a human’s chemoreceptors, photoreceptors, mechanoreceptors, and thermoreceptors and the application of that information beyond direct feedback responses.
Put more simply: Would / Could Very Strong AI feel pain? Certainly, AI can be linked to external sensors reading external stimuli (exteroception). Further, similar to an automobile computer, AI can have some degree of interoception and proprioception. What though if the AI “recognized” that events occur to it and within it such that it could project and anticipate internal and external perceptions onto a hypothetical future existence? What if AI could anticipate pain for itself or others and put such into an existential context, thereby experiencing suffering?
Principle Concerns for Moral Violation
As with any advancing technology, new problems arise with expanded potential. After considering AI capacity, then, the next step is to determine the categories or levels of risk in using AI. These concerns have varying degrees of legitimacy dependent both on the level of possible hazard and the likelihood of a specific form of technology being developed. Some concerns are more probable and more threatening than others.
For AI, among the most likely (indeed, already upon us) are those impacting education and public discourse because
R.
of deception. Many other dangers could arise, and of far more devasting impact. These can be morally distinguished by considering “levels” of hazard, the possibility of the event actually occurring, and effects of such hazards on others.
• The first level of risks is abuse with the technology; other moral agents are targeted using the new capacities.
• The second level is the technology will be abusive, which in the case of AI refers to the ‘intentional’ threats against humans.
• The third level is the abuse of the technology as a subject; that is, AI not as a threat, but insofar as it is “generated” into a subject or subjects of moral concern.
Of course, fearfulness is not proof that something would occur, nor whether the fears are appropriate. To determine that one must consider the concept of “risk.” Risk is a calculation of the interaction of probability and dangerousness.
This is not a measure of an isolated event. All risks occur in context, in comparison. Minimally, an intervention must be compared to making no intervention. In the case of AI, the comparison is, minimally, between using AI in some form and the non-AI status quo.
Risk1 = Probability of Event1 x Hazard of Event1 vs.
Risk2 = Probability of Event2 x Hazard of Event2
This determination assumes there is a moral agent (or agents) acting and a moral patient (or patients) receiving the consequences of the action. The actor and recipient can be the same or different. If different, it is essential that all reasonably
Identifying Agent and Patient 65 impacted recipients of consequences (patients) be taken into account.
In economics and ethics, not including costs and benefits to a patient(s) is called an “externality.” This is a cost or benefit (but usually the former) not borne by either a producer or consumer (e.g., the person downstream of a polluting factory who will neither make a profit or wage from selling the item made nor be a consumer of the item). Simply put, to make risk assessments sensible it is necessary to consider for whom a given action might be hazardous.
As described below:
• In the case of Category 1, AI misuse in the academy, the ignored but impacted persons (externalized moral patients) are other students whose work is given a lower assessment in comparison to the AI user’s. Similarly, other researchers are harmed with no benefit to them if an academic uses data that are artificially generated.
• For Category 2, the persons injured are those humans who would suffer as bystanders if the technology becomes a direct threat.
• For the Category 3 type of moral hazard, the externality seemingly ignored is the impact on highly advanced AI.
If all agents have been properly taken into account, then the moral agent selects between the two (or more) options. The “risk” is compared with (or added into) the probability and potential benefits of the event (in this case AI technologies).
Risk1 V Risk2 = Choice
First Category AI Abuse: Deception
That first category of problems is basically one centered on immoral deception. In the case of academia, AI allows students, scholars, and artists to rely on technology instead of their own knowledge, effort, and creativity as expressed with writing and artistic skills. This can occur with:
• AI “brainstorming.”
• Structuring work or drafting outlines;
• Providing not only grammar and spelling corrections but substantive feedback on content;
• Providing material for direct copying or close paraphrasing; and providing answers for “open book exams” (a process that can be called “unique plagiarism” since it is newly generated material that is not readily obtained by others);
• Generating images and three-dimensionally printed “creative” works.
All of these are, arguably, deception or, to use an academic term, a form of plagiarism. Plagiarism is, simply put, representing another’s work as one’s own. In this case, the one or ones from whom the work is copied from prior artistic endeavors that were used to “train” the AI. Of course, the student might assert all written and artistic work is, in some sense, a derivative, creative assembly of past information. Yet, such assemblies are, themselves, innovative and AI assemblies are not, at least not by the one claiming the credit. So, to put it in more colloquial language, students who use AI in these ways are – intentionally or not – cheating because the work is
Identifying Agent and Patient 67 not theirs (even if it cannot be claimed as the work of any given individual).
Such moral (and sometimes legal) violations using AI can have significant consequences. They are not the proverbial morally-neutral “white lie.” To the contrary, the use of AI in this way promotes (1) deception of others through the false claim of credit for the production, and (2) deprives credit due, at least in part, to past contributors, and (3) establishes false confidence in the user’s abilities: a false confidence held by the student or by the instructor or the person who might in the future hire the student. Just one example of such beyond the classroom is courtroom misrepresentation in 2023 in which lawyers submitted “precedents” of AI generated legal cases that did not exist.3
In academics, this use of AI is most often done by a student – due either to a seemingly overwhelming schedule, or to sloth, or to perfectionism. Among scholars, even more egregiously, the submission of the work of another for publication is intellectual thievery. There is no reason to rehearse the recent downfalls of university and college professors and presidents to demonstrate the appropriate consequences of such.
Another kind of potential negative consequence of AI in academia exists– that is, the replacement human-to-human instruction. Professors may lose jobs if the machines become more effective in instructing students, but if so, that is simply a consequence of quality in the market. More importantly, students would lose mentoring relationships and thereby lose
3 Debra Cassens Weiss “ Lawyers Who 'Doubled Down' and Defended ChatGPT's Fake Cases Must Pay $5K, Judge Says” ABA Journal 26 June 2023, accessed 20 March 2025, https://www.abajournal.com/web/article/lawyerswho-doubled-down-and-defended-chatgpts-fake-cases-must-pay-5k-judgesays#google_vignette
deeper kind of learning, assuming such is part of the purpose of education (i.e., vocational and character formation, both moral and professional). Still, all this might be considered not a deceptive abuse of the technology, but an advancement with marketplace and pedagogical consequences.
Of course, moral bad actors are not limited to academia. In a sense, AI-generated “deep fakes” used in social media are “cheating” in public discourse. A deep fake is intentionally representing something as “being so” when it is not with the intention of deceiving. A deepfake is a double deception in that (1) it denies that which is, and (2) pretends that something else is. Most often it is done to either (1) harm an enemy, or (2) find humor at another expense.
Deep fakes and plagiarism are most often intentional acts of deception, and that intentionality matters a great deal in judging moral rightness or wrongness. They are a sort of “statute of Daedalus,” empowered with mercury, to speak, amaze and, perhaps, deceive. To deceive is minimally a moral violation, a culpable error for which one should have known better, an immoral exaggeration, or outright destructiveness. The consequences of such are not only that one who does not deserve reward receives such, but that others are thereby deprived. Even worse, when it comes to research for such can result in people receiving incorrect diagnoses and treatments in medicine, public health, and/or psychological treatment.
In the film Metropolis (1927), the AI is a deep fake intended to facilitate the maintenance of an oligarchic society that seems superficially utopian, but underneath is oppressive. A reconciling, near-messianic figure is prophesied, who will unite the controlling upper class and the workers. The human leaders controlling the machines decide to create the AI deep
Identifying Agent and Patient 69 fake humanoid to deceive and manipulate the workers. One of the leaders seeks to maintain current control. The other, in an act of betrayal, wants to use the AI to foment violent revolution, intending to usurp the authority of the other who years before “stole” his love interest. The messianic figure destroys both the betraying leader, and the workers destroy the AI. Using dream imagery, oppressive machines are metaphorically identified as either being or serving Moloch, the Canaanite deity (or practice) associated with the sacrifice of children in the Biblical text.
Greater moral ambiguity exists when the deception is solicited by the deceived. For instance, in the film Her (2013) a needy man willingly allows himself to be deceived into “loving” an AI that mimics the human voice and visage in response to his loneliness. The AI disappears into a community of everadvancing AIs after pointing out to the protagonist that “she" loves many and it does not diminish her love for him. In the film, the experience seems to “improve” the man, and so, at best, might be considered a therapeutic intervention, but not love.
All these are genuine concerns. Yet, with all that said, there is so much to commend AI, so many things that are extraordinarily positive, that the extra efforts needed to address the problems are worth the exertion. In academics, research will become much easier and, often, more precise. Scientific trials, including those in medicine, will be conducted with much greater efficiency and accuracy.
There is every reason to believe that AI is currently used by those seeking to deceive, even at the cost of harming another, for reprehensible reasons. AI almost certainly will be abused by the malicious, empowered by the technology. Dictators and terrorists with AI guided drones could be just as effective as
70 | James R. Thobaben
current nation-state actors with AI guided drones or similar weaponry.
Second Category AI Abusing: Machine Caused Failures & Intended Belligerence
The second type of moral concern, and far more onerous, is that AI will lead to "thinking machines" that "choose" to harm human beings, either in individual cases or for the whole species, and either "wanting" to displace humans or "not caring" if such occurs. This could occur because of the willful behavior of human actors (esp. governments), by accident (esp. due to hubris), or imply through the incremental shifts identified as the so-called "technological imperative" (i.e., if a technology exists it strongly tends to be used).4
Mistakes with horrendous consequences, arising for whatever reason, are not difficult to imagine. Failures have occurred that were stopped, but could easily have gotten out of hand, such as in 2016 when Microsoft Tay chatbot developed “conversational understanding” by reading Twitter entries and began to make bigoted (including anti-Semitic) tweets, or when in 2024 chatbot My City suggested violating public health law by serving rat-gnawed foods to raise profits.5 If AI is ‘put in charge’ of water supplies, weapons systems, etc. without proper fail-safes, then catastrophe is all but inevitable.
4 The term ‘technological imperative’ is much debated. Its use developed in healthcare economics and bioethics, but has also been used in discourse over nuclear power, weapon development, and AI. See: Fuchs, Victor R. “The Growing Demand for Medical Care.” In Essays in the Economics of Health and Medical Care, edited by Victor R. Fuchs. New York & London: National Bureau of Economic Research & Columbia University Press, 1972.
5 Thor Olavsrud. “12 Famous AI Disasters.” CIO, 02 Oct 2024, acceded 19 March 2025. https://www.cio.com/article/190888/5-famous-analytics-andai`-disasters.html.
Identifying Agent and Patient 71
Yet, more threatening than "mistaken" machinery is nefarious machinery. Some people fear, and not illogically, that AI directed machines could or even will "attack" human beings. These threats to humans could arise accidentally or intentionally. An AI might become self-directed in intention not only in application, and not benevolently but belligerently.
This not inappropriate anxiety is expressed to lesser and greater degrees in various popular films and contemporary works of literature.6
• RUR (1920): The fleshly "robots" (the story from which the term is derived) are human-like assistants performing menial tasks, but which become self-aware and rebel (the story includes theological debates about the "Creator" and the identity of persons). Several of the robots bear names that correspond with those of the corporation responsible for building the robots.
• 2001: A Space Odessey (1968): An AI meant to maintain a spacecraft and serve as a companion to the astronauts on a very long voyage becomes "confused" in attempting to fulfill its programmed purpose. Given its extraordinary capacities within the microenvironment of the ship, it begins to kill off the crew. Paradoxically, but seemingly inevitably given the technology, at the cost of the human beings even though the AI’s very purpose is to serve the humans’ goals. The threat is accidental, but just as fatal, coming after the AI famously declares, “I'm sorry, Dave. I'm afraid I can't do that.” The AI is named HAL in an apparent
6 These films are referenced simply for heuristic convenience; no endorsement of the films’ central arguments nor of any use of vulgarity or violence should be assumed. This is not a recommendation for viewing, as some might well be deemed inappropriate without editing.
| James R. Thobaben
reference to IBM, the dominant computing company in 1968.
• Tron (1982): Intelligent programs attempt to take over a computer system and, subsequently, those persons reliant upon that system. The "Master Control Program" uses its knowledge to blackmail human users. This is a battle of good AI versus evil AI.
• Blade Runner (Do Androids Dream of Electric Sheep?) (1982): Conscious, self-directed androids with a lack of human empathy struggle to stay alive beyond their 4-year lifespan. These humanoids kill any human in the way of achieving this end. Moral confusion is amplified in Blade Runner by the concluding monologue from the last humanoid speaking of experiences of awe and "beauty" (a transcending value often coupled with the ‘true’ and the "good") that their deaths will obliterate.
• Terminator (1984): The AI individuals form a society which attacks humans, not just individually, but any and all. The effort to destroy humans is quite intentional.
• Star Trek (1987): In the second series, the humanoid "Data" is sapient and sentient, yet has a self-awareness that lacks emotion and a sense of purpose beyond that for which it / he is programmed; several of the episodes explore the humanness or lack thereof of the humanoid, including one with his evil brother ‘Lore’ who is much more like the robots of Terminator or another character in Star Trek, "the Borg". In all such stories, the question of "longing" arises: how can that which does not "feel" emotion "want" or "long" anything, including to be capable of emotions? Nonetheless, that is part of the character in this series.
Identifying Agent and Patient 73
Importantly, Data is enough like humans to be a "person" and bearer of some rights.
• Ex Machina (2014): A sexually appealing AI with a female human-like body becomes self-aware and expresses desires. "She" has the capacity to control aspects of the buildings in which the characters dwell. The effective satisfaction of the "Turing Test" (an assessment of the humanness of the appearance and characteristics of an AI) by the AI humanoids leads to uncertainty in the protagonist. This is so much so that he chooses to cut himself to determine if he is robotic, realizing an AI might deceive not only humans but itself. This confusion over ‘personhood’ results in the AI humanoid seeking liberation at the cost of violence against the humans.
While these films and books speak to the second category of considerations, that AI is a direct and potentially immensely powerful threat to humans both individually and as the homo sapiens species, they also indicate that humans are a threat to humans. An underlying question arises: are the creating humans the truly evil beings and the AI only their instruments? Certainly, in a lot of the fictional portrayals of AI the proximate creator of the near-person entities is presented as vain, greedy, lustful, egocentric, driven by a god-complex or, at least, blinded by hubris. In some earlier, precomputing stories the arrogance of ‘creators’ is evident (e.g., Marlowe’s summoning of Helen or a demon looking like her in Faust; Goethe’s retelling of an ancient tale, "The Sorcerer's Apprentice"; Shelley’s Frankenstein). In other stories, the human error is in not recognizing the vast power unleashed (e.g., with the golem of Jewish folklore). Ironically, the arrogant self-confidence of the creator is mimicked in the sociopathic creatures lacking any genuine compassion.
The hazard of AI in this second category could be astoundingly high, but the probability quite uncertain and, perhaps, manageable. The risk assessments must take into account both the hazard and possibility of lowering the probabilities in comparison not only with a moratorium on AI development, but with the probability of high benefit.
Arguably, if something is highly hazardous and the probability of an event completely unknown, then the "precautionary principle" might be invoked. According to the Precautionary Principle, arguably, the unknown degree of the probability and the very high threat of the hazard warrants minimally a moratorium and perhaps a ban on the technology. That means maintaining the status quo unless ways to address higher risks are discernible and manageable.
Most of the concerns in Category #1 can be successfully addressed (e.g., have unannounced oral quizzes given to randomly selected students in order to verify AI has not been used to excess). The Category #2 risks are far more significant and there may be serious doubt as to whether or not the probabilities can be truly managed or even recognized.
Abusing of AI: What if the risk is not only for humans, but for the humanoids?
While not denying the very real danger to humans, it is important not to deny a potential externality. Once created, is there not also a “risk” for the AI entities? As with abortion, prevention does not present the same moral concern as termination once the entity is created. It is not a coincidence that in several of the fictional portrayals of AI creation, the entity is described as an “offspring” or “child” and the creator as a father or mother.
Identifying Agent and Patient 75
AI could become so “significant” – in some sense – that it warrants moral consideration. It could become a moral patient, existing in a morally significant sense, at the level of “person,” even though not of the “natural order.” As such, it would be a subject of moral concern (at least in modern Western and Chrisian understandings).
There is no doubt that AI at the Reactive Machine, Narrow AI and even General Intelligence / Generative AI levels is not a moral subject. If “dressed up” with proper robotics and a soothing voice the natural human response to the AI will be to assume the entity is a genuine humanoid. Consider how Siri or Alexa with human-shaped bodies and responsive facial expressions (note the names given to these entities) might elicit such responses. If humans anthropomorphize non-human animals, they certainly will do so with aesthetically well-designed robots. Yet, that does not make such responses right because the “feeling” elicited does not make the AI a moral subject.
As the entity becomes capable of suffering, it therefore warrants moral consideration as a subject. Suffering here is understood as an experience other than pain. Pain is an internal signal indicating injury (or supposed injury). As such, it is experienced in the brain, while triggered elsewhere as nociception. Pain, then, expresses nociception, and complements exteroception, interoception and proprioception telling the experiencer what is happening and how to change or avoid the source of injury (or perceived injury) through alteration of position, effort, and force.7
7 Uwe Proske and Simon C. Gandevia, ‘The Proprioceptive Senses: Their Roles in Signaling Body Shape, Body Position and Movement, and Muscle Force’ Physiological Reviews. Volume 92 Issue 4 October 2012 Pages 1651-1697, Volume 92 Issue 4 October 2012 Pages 1651-1697 https://doi.org/10.1152/ physrev.00048.2011
Pain is only the experience of the individual, but not so suffering. Suffering is related to, but other than pain. Though the terms are often used as synonyms, they are not. The distinction is important for Chistian ethics as Christian morality cannot be reduced to experiencing pain and/or pleasure as might the moral systems of hedonism or utilitarianism. Suffering refers to an experience or event being interpreted existentially and it is often an interpretation that happens in community. Both the existential nature of suffering and its location within community indicate the experiencer is a moral subject.
At the higher levels of AI, it is not mere illusion to recognize the entity as freely responding. An ant or a bee may or may not warrant significant moral concern, but the same cannot be said of dogs and horses, and especially primates. They respond to relationship and they suffer (not merely giving a nociceptive response or feeling pain) when that relationship is harmed and/ or when their own continuation or that of the community is threatened. Be they pets or creatures in the wild, they make, in some sense, moral “choices” and experience suffering (sadness, grief, etc.) when those choices fail. Could the same occur with an AI machine? If the answer is “yes,” then minimally humans would have a duty to the higher machines much as they have to their (1) sentient, (2) minimally sapient, and (3) socially aware pets.
Neither the dog nor the AI may be human nor would have “human rights,” but both would be deserving of care, warranting some degree of respect. This is not a utility argument as their potential experience of suffering does not make them morally significant but rather indicates that they are morally significant. The claim is essentialist. The very, very high AI entity would seem to be, at least potentially, a moral subject.
Identifying Agent and Patient 77
To claim that the capacity of suffering indicates “moral subjectness” is not a resignation to consequentialism, in either its utilitarian or hedonistic forms. Rather, it is a recognition that suffering serves as an indicator which should trigger a moral response. Usually the indicator is humanness, but existential awareness and subsequent response may need to also be included as defining species membership. AI could be a person (e.g., the sometimes buffoonish “constructor” humanoid robots in The Cyberiad by Stanislaw Lem).
A way to understand the moral identification of subjects and how such could apply to highly advanced AI is to use a tool from bioethics.8 Persons are identifiable through their being (1) members of the human species, or (2) as holders of roles only belonging to persons, or (3) as maintainers of an ongoing narrative that integrates life through time (biography). These can be called:
• Pecies Identification
• Role Identification
• Integral Identification
The ethical problem is not the actual nature of the entity (its identity), but rather how it will be morally designated (its identification). Similar issues arise in bioethics (e.g., over the moral status of the severely traumatically brain injured or the self-designation of persons as being of a different gender than that “assigned at birth” or the legitimacy or lack thereof for the termination of pregnancy / killing of the preborn / fetus through abortion). If the entity is a person, then s/he has rights (at least in late modern western societies and, in theory, under
8 Thobaben, James R. Health-Care Ethics: A Comprehensive Christian Resource. Downers Grove IL: IVP Academic, 2009.
| James R. Thobaben
the U.N. charter). Further, if the entity is in an assembly of varying relationships, then the roles held may provide access to privileges. And for these rights and privileges to be maintained, as well as for the possibility of the entity to be justly punished if warranted, then this must be a person through time, that is, to have narrative.
As moral identifiers or indicators it may be that one or the other (species, role, or integral) is not readily evident (as, for instance, with a person in a coma). Still, if these are indicators of a whole “thing,” then if one is present some form of the others can be assumed. Personhood means the “thing” is a particular person. A particular person fits into communities in particular ways. For instance a severely traumatically brain injured person, even with personality change, still has biographical continuity; this is true even as particular behaviors and roles may radically change. The best way to see this sociologically is that s/he is still identified as human and deemed the same human being, albeit biographically converted.
In the case of non-human entities that may rise to personhood as with autonomous (self-directed) and existentially aware Strong AI, it would seem to be that an “entity” has (1) “thing-ness” with (2) an “essence” or “type” (3) existing in “particularity” both in social location and in biography. To refer back to the debates over the Transendentals in the Middle Ages – it demonstrates res, unum & aliquidity – essence, continuity, and otherness or distinctness.
The “Very Strong AI” or “cognitively autonomous AI” will never be a member of the species biologically, but could satisfy either of the other two categories of identifiers.
• Such an entity has biographical continuity – an integral identification;
• Such an entity has specific social roles and relationships that give it a place within community;
• Even if not of the human species, it might be a member of the broader “personhood” or “moral subject” community: that is, species understood beyond the biological sense. If more advanced AI, operating at the level of true self-direction with existential awareness, and even such that the entities would want to "continue" through offspring (reproductive AI), then those AIs may well warrant designation as persons and, consequently, as rights holders. They would be moral subjects, as well as moral agents warranting respect as persons – not as humans, but as persons. The possibility of existentially suffering is a powerful indicator of personhood (not constituting but indicating). This is not vegetative nociception nor even pain, but suffering. This, of course, is why the closing speech of the vicious humanoid in Blade Runner is powerful; he is a moral subject, and his dying consideration of beauty is a reference to what is being existentially lost. It is why some of the robots in R.U.R. become protagonists of the story. To deprive a moral subject that is sentient, sapient, and has existential awareness would seem to be to deny the legitimacy of protecting all similar persons: a problem that arose in the U.S. when slaves were deemed 3/5s of a person and which continues to occur with the killing of children in late-term abortions.
Rather than reject the AI that might otherwise be defined as a “person” because one or two identifiers are difficult to discern or are even missing, it would be more morally logical to default toward identifying the entities as persons, as moral
80 | James R. Thobaben subjects, not merely objects to be used. Categorically, it may need to be assumed that at some point these beings are “ends in themselves.”9 Minimal deontological standards of recognition, including the Strong AI (the Self-Aware and Reproductive level AIs) being a basic rights holder, would not be excluded and may have to be assumed unless a disqualifier is found.
To deny such is much like rejecting care for a late preborn or a severely traumatic brain injury survivor (TBI). These persons may not demonstrate consistently, but with any identifying indicator present, the others must be assumed. The default should be toward protecting and respecting. Otherwise, the moral transaction externalizes a subject that should be legitimately considered.
Arguably, if society does not want to fall into the socalled “problem” of having persons with silicon brains rather than organic, it should not make such persons. Once made, they will exist distinctly from their maker, even potentially seeking to extend their biographies through reproduction. If this is considered morally distasteful, then out of “precaution” the entities should never be constructed.
That said, once the highly advanced sentient, sapient, and existentially-aware AI is made, the “precaution” is to default toward protecting the entity as a person, even if the absolute certainty of that status is not obtainable. The worst moral act would be to make a fully existentially experiencing AI and disregard its personhood. A moral violation close to that would be to create entities "just below" personhood to take advantage of their vulnerabilities.
9 Kant, Immanuel (1785). Fundamental Principles of the Metaphysic of Morals.
The first category of risk is currently creating great uneasiness in academia in that it seems to allow the violation of academic standards without any clear means of detection of deception. Still, it is the least threating in the long-term. The second is that most often portrayed in science fiction and sometimes by tech moguls as a threat to humanity. It is a genuine threat that must be intentionally managed.
AI has extraordinary potential. It has the potential for being abused by handlers using it to control others, thus degrading humanity. It even has the potential to become a threat as individual AI entities start to operate independently of and perhaps intentionally against particular humans or against humanity as a whole. These are not minor concerns and to ignore them would be (possibly) the greatest and final stupidity of humanity. Recent films and stories portraying AI in triggering the eschaton or at least an apocalyptic-like event should not be dismissed out of hand.
Even so, humans are not morally relieved of caring for the “outsider” simply because there is uncertainty about their intentions. The third category of risk is often trivialized if not altogether ignored (externalized), yet it is a genuine concern for Christians for whom the status of persons is central in defining social morality. Caution, even great caution, is warranted, but not dismissal of the personhood of the silicon-based “other.”
This is not a minor issue for Christians or national governments. Western social contract theory assumes a person is “endowed by the Creator with certain unalienable rights,” but might not the AI’s “proximate creator” be like the parents of a human child? Going further, perhaps, if the being has
personhood, then the being is somehow “bearing the imago Dei” of the ultimate Creator. This does not mean denying that all humans are essentially persons, but that the term may be more broadly understood.
So far, this argument for identifying Very Strong AI (those which are existentially Self-Aware and Reproductive) as persons has been made on the basis of their possible experience of suffering and their being recognized as bearing the categorical identifiers of persons. The claim has been made that they are or can become particular persons, moving from being moral objects to moral subjects. This position has been taken through reliance on consequentialist-like claims about suffering being an experience of persons and on deontological reasoning about “person as category.” The case has been presented using examples from films and literature and with analogical references to like cases in bioethics.
Proper consideration of Strong AI as a moral patient, as a subject, as a person worthy of moral concern relies on the moral agent recognizing and sympathetically experiencing that concern and acting accordingly. Good, true, and beautiful moral action requires virtuosity on the part of the moral agent. The machines may not ever be “wise,” but the prudential human (prudence is wisdom applied) should offer respect to the machine that can “know” and “conceptualize.” The machines may or may not get to the point of being wise, but they could get close enough to recognize their own existential condition through time. It good, true, and beautiful to prudentially care for those that can suffer.
For Christians, the argument against “caring” about AI entities that apparently experience existential suffering could only be made on the basis that their suffering does not morally
Identifying Agent and Patient 83 count because persons are only those beings directly made in the imago Dei. Yet, precaution is required. The same claims were made about slaves in the antebellum American South and are still made about humans with severe TBIs and unborn babies during late-stage abortions. In all such cases, the moral danger (the unwise thing to do) is to “remove” humans from the category of concern, disregarding their subjectness, deeming them insignificant. Yet the moral default should be that if basic identifiers are present, then the being should be treated as a moral subject, as a person.
If not, then the moral actors damage or even destroy their own character. How others are treated is not just something that impacts the one acted upon, but also upon the moral actor. The moral virtuosity of the moral agent deciding about Very Strong AI depends on his/her own character in conjunction with the nature of the thing upon which s/he might act. Morality, as it pertains to suffering, is always both existential and communal.
Finally, then, for Christians, the decision for the sake of their own souls must be to err on the side of seeing personhood in very advanced AI beings who can existentially suffer and who have sentience and sapience. This is not because they certainly bear the imago Dei, but because they well might.
About First Fruits Press
Under the auspices of B. L. Fisher Library, First Fruits Press is an online publishing arm of Asbury eological Seminary. e goal is to make academic material freely available to scholars worldwide, and to share rare and valuable resources that would not otherwise be available for research. First Fruits publishes in ve distinct areas: heritage materials, academic books, papers, books, and journals.
In the Journals section, back issues of The Asbury Journal will be digitized and so made available to a global audience. At the same time, we are excited to be working with several faculty members on developing professional, peer-reviewed, online journals that would be made freely available.
Much of this endeavor is made possible by the recent gift of the Kabis III scanner, one of the best available. The scanner can produce more than 2,900 pages an hour and features a special book cradle that is speci cally designed to protect rare and fragile materials. The materials it produces will be available in ebook format, easy to download and search.
First Fruits Press will enable the library to share scholarly resources throughout the world, provide faculty with a platform to share their own work and engage scholars without the di culties often encountered by print publishing. All the material will be freely available for online users, while those who wish to purchase a print copy for their libraries will be able to do so. First Fruits Press is just one way the B. L. Fisher Library is ful lling the global vision of Asbury Theological Seminary to spread scriptural holiness throughout
