Enhancing Learning and Teaching at CPUT through Artificial Intelligence: A Guide

Enhancing Learning and Teaching at CPUT through Artificial Intelligence: A Guide

We extend our gratitude to the contributors who have lent their expertise and insights to the development of this guide:

Dr Faiq Waghid, Head of the Department for Academic Staff Development, Fundani

Dr Xena Cupido, Director, Fundani

Dr Browyn Swart, Learning and Teaching Coordinator, Faculty of Engineering and the Built Environment

Assoc. Prof. Muhammad Nakhooda, Learning and Teaching Coordinator, Faculty of Applied Sciences

Assoc. Prof. Zayd Waghid, Faculty of Education

Assoc. Prof. Lalini Reddy, Director, Centre for Community Engagement and Wil

Xolani “MAMPHITSHA” Booi, Education and Transformation, CSRC

Ekaterina Rzyankina, Department of Maritime Studies, Faculty of Engineering and the Built Environment

Thandiwe Sidzumo, Graphic Designer, Fundani

Images & Illustrations by Dall.E 3

1. Introduction

The purpose of this guide is to provide staff at Cape Peninsula University of Technology (CPUT) with practical and ethical guidance on incorporating generative AI into their learning, teaching, and assessment practices. Generative AI was used to write parts of this guide. The specific prompts used to develop this guide were formulated with the unique context of CPUT in mind and were shaped by the collective input of a deliberative task team. This document acknowledges the potential impact of generative AI on pedagogy, curriculum design, and academic integrity. It advocates for a balanced and open-minded stance towards learning and teaching, bearing in mind the evolving landscape of knowledge production and distribution in the age of generative AI. Given the spectrum of reactions it may provoke, the guide advocates for the need to foster a continuous discourse which clearly outlines the benefits and challenges tied to artificial intelligence. This serves as an initial guide for the process of exploring and purposefully integrating generative AI, specifically within the context of the Cape Peninsula University of Technology.

At the risk of duplicating very pertinent policies in this guideline, a framework of policies associated with this guideline is listed in the diagram. It is important to read this guideline in conjunction with those policies.

Learning and Teaching at CPUT through Artificial Intelligence: A Guide

Although the guide adopts a positive stance on the integration of generative AI into higher education, it simultaneously recognises the series of reservations outlined in the table that follows. These reservations span from ethical dilemmas to concerns over content quality, the reinforcement of biases, intellectual property, an over-dependence on technology, the shortfall in personalisation, data privacy concerns, the implications for cost and accessibility, the complexities involved in evaluation, and the danger of spreading misinformation. By acknowledging these potential issues, the guide does not seek to diminish their importance but rather to accentuate their essential role in the discussion about generative AI in higher education. In light of the varied responses that the adoption of generative AI may rouse, the guide champions the imperative of maintaining an ongoing dialogue that thoroughly articulates the benefits and acknowledges the challenges tied to the use of artificial intelligence. This balanced approach supports a thoughtful and informed integration of generative AI, striving to harness its strengths while vigilantly navigating its challenges.

Reason for not using generative AI

Ethical Concerns

Quality of Content

Bias and Fairness

Intellectual Property Issues

Dependence on Technology

Lack of Personalization

Explanation

The potential for promoting academic dishonesty, such as plagiarism.

Generative AI-generated content may lack the depth, nuance, and critical thinking required by higher education.

Generative AI models can inherit biases present in their training data, leading to outputs that could perpetuate stereotypes or unfair representations.

The use of generative AI raises questions about the ownership of generative AI-generated.

Overreliance on generative AI for content generation and educational tasks could diminish students’ and educators’ skills development.

While generative AI can produce generalised content, it may not adequately address the specific needs, backgrounds, and learning needs of individual students.

Data Privacy Concerns

Cost and Accessibility

Evaluation Challenges

Misinformation Risk

The use of generative AI in education requires significant data input, raising concerns about the protection of sensitive information.

Implementing generative AI solutions can be costly, potentially widening the digital divide and affecting equitable access to educational resources.

Assessing generative AI-generated work can be difficult, as it may not reflect the student’s own knowledge, understanding, or critical thinking abilities.

Generative AI has the potential to generate content that is factually incorrect or misleading, hampering the learning process.

2. Generative AI in the Context of Higher Education

“Generative artificial intelligence models are computational systems trained on large datasets to generate human-like responses or output. These models can assist with a wide range of queries and provide information across multiple domains” (OpenAI, 2022). Generative artificial intelligence (AI), exemplified by Large Language Models (LLMs) such as OpenAI’s ChatGPT and Google’s Gemini, have the capability to generate a wide range of outputs, spanning textual compositions, programming code, musical compositions, and visual imagery from natural language prompts. A plethora of research highlighting the potential of generative AI to improve learning and teaching in various knowledge disciplines already exists. However, despite the many perceived benefits of AI, its use in higher education necessitates an exploration of the potential ethical challenges and reservations pertaining to academic integrity.

In recent years, the higher education sector has seen several disruptions. Between 2015 and 2017, there was student protest in response to the hike in student tuition fees and protest calling for the decolonisation of university curricula. These disruptions played a role in prompting transformations in curriculum and programme development, in addition to fostering considerable scholarly investigations into the decolonisation of university curricula. Moreover, the COVID-19 pandemic, from March 2020 to early 2022, presented a significant disruption. In these contexts, many universities faced difficulties in the process of rapidly transitioning to Emergency remote teaching. The disruptions and ensuing problems have led to a reconsideration of how the curriculum is delivered and have resulted in the introduction of flexible approaches to learning and teaching in universities.

The introduction of generative AI can be considered by some to be yet another disruption in learning and teaching in higher education, one requiring careful navigation. According to Rousseau (2023), the university’s longstanding and uncontested dominant status in the production and dissemination of knowledge has now been called into question due to the progress of digital technology and generative AI. This has led to a range of different responses from universities around the world regarding the incorporation of generative AI tools within the context of learning and teaching. Certain universities have implemented policies that explicitly prohibit the utilisation of such tools. These institutions express their apprehension regarding issues of academic integrity and the possible suppression of independent thought and creativity among both students and teachers. Conversely, there are higher education institutions that view these tools as advantageous to supporting the generation of ideas and to enriching discussions inside the context of learning and teaching. The Cape Peninsula University of Technology (CPUT), with its forward-looking Vision 2030, recognises the immense potential and transformative power of generative AI in shaping the future of education. As we move towards a world increasingly intertwined with technology, it is imperative that guidelines, such as those presented in this document, are established to ensure the ethical, responsible, and effective use of generative AI. It is envisaged that, by means of these guidelines, CPUT not only positions itself at the forefront of educational innovation, but also ensures that its students are equipped with the knowledge and skills to navigate an AI-augmented world.

3. Navigating the Guideline

Not to overload this guide with content with numerous examples showcasing every aspect of generative AI functionality, useful examples of prompts are presented instead. This approach enables users to input these prompts into the generative AI platform of their choice, exploring the functionality first-hand. By presenting the information in this way, we strive to provide a succinct and pragmatic guide, prioritising its readability and usefulness.

4: Bing AI interface: A sophisticated chat tool accepting a range of inputs from text and images to PDFs, Word Documents, and Mathematical Expressions.

4. General Recommendations and Principles for the Use of Generative AI

The section below offers general recommendations and principles for the responsible, informed, and effective use of generative AI in various learning and teaching contexts at CPUT.

Clear rationale

The inclusion of any digital technology in the learning and teaching context should be undertaken with thoughtful consideration aimed at enhancing the pedagogical process. The integration of generative AI in this process needs to be attuned to this consideration.

Clear instructions

When suggesting the use of generative AI in a learning and teaching context, clear instructions for its use should be provided. The aim of these instructions should be to develop a clear understanding of how to interpret the results and how to use the technology in informed and effective ways to facilitate learning while at all times being mindful of potential challenges and limitations.

Student comprehension

It is important to develop in students an understanding of the functioning of generative AI. It is crucial for them to grasp that generative AI functions by generating word sequences that are likely based on its training data. It does not possess any form of comprehension, cognition, or reasoning abilities.

Critical Thinking

Critical thinking is the disciplined process of actively and skilfully conceptualising, applying, analysing, synthesising, and evaluating information gathered from, or generated by, observation, experience, reflection, reasoning, or communication, as a guide to belief and action (Paul & Elder, 2006). It is imperative to underscore the significance of critical thinking when analysing the role of generative AI. While generative AI has the potential to offer valuable insights, it should not be considered a substitute for the fundamental cognitive process of critical thinking (Paul & Elder, 2006). The cultivation of graduates who possess the ability to exhibit autonomous thinking and generate innovative ideas is a fundamental objective of CPUT Vision 2030, and of higher education.

OPENNESS

Corroborating information

Corroborating information generated from AI by curating other reliable sources ensures the accuracy and reliability of the information. Generative AI tools should together serve as a complementary source of information rather than constituting the sole reference.

Openness

Individuals should be encouraged to state whether and in what contexts they have used generative AI. This means highlighting instances of the use of tools such as language models when completing tasks.

Rigorous discourse

Discussions pertaining to the ethical aspects associated with the utilisation of generative AI should be encouraged. These should relate to its constraints, inherent biases, copyright implications, labour and environmental considerations, as well as data ownership rights. In the context of these discussions, it is crucial to engage in a rigorous assessment of the generative AI’s outputs in order to ensure adherence to academic integrity norms.

Adherence to these recommendations and principles may assist CPUT staff and students to effectively integrate the use of generative AI. These recommendations and principles recognise the need for fostering critical thinking skills, maintaining a clear educational rationale, addressing issues of ethics, and for advocating transparency. These considerations are essential within the broader context of the ever-evolving higher education sphere.

5. Getting the Best Out of generative AI: Tips for Engineering Prompts

The guidelines below are intended to assist CPUT staff and students in navigating their interactions with generative AI with the purpose of obtaining reliable, relevant, and concise information while also being cognisant of the potential limitations.

Suggestion Description

Be Specific

Being specific and clear with prompts enhances response accuracy.

Sequential Questions Broad topics benefit from breakdown into sequential or related questions, enabling more focused and detailed answers.

Specify the Desired Specifying the desired format, such as a list, paragraph, or Format summary, refines the response.

Anticipate Bias

While generative AI aims for neutrality factuality, considerations about and Ambiguity potential bias or balanced coverage might include prompts like “Provide a balanced overview of the pros and cons of nuclear energy.”

Utilise the Knowledge Training data extends up to April 2023. Queries about post-cut-off Cutoff events or information may have potential gaps in response.

Ask for Sources For evidence-backed information, prompts might seek the basis or or Bases sources for generative AI statements, e.g., “Evidence supporting the health benefits of meditation?”

Refinement Follow Ups

Refining follow-up questions based on initial answers facilitates an iterative approach to obtain clearer information.

Limit Information Narrow, targeted queries often prove more effective than broad, openOverload ended ones.

Leverage Multimodal Including images or non-textual prompts can add context or cater to Capabalities visual based queries.

Stay Updated on Generative AI tools’ capabilities evolve. Periodic checks for new features and Capabalities guidelines are recommended

6. Recommendations for Privacy and Security of Data

According to the guidelines and practices of the currently-available LLM platforms, privacy and security of data are top priorities. The following recommendations, taken from OpenAI in particular, may assist in maintaining the privacy and security of data.

Suggestion Description

No Personal Data OpenAI retains customer API data for 30 days but does not use Retention this data to enhance its models.

Anonymity Users are advised not to disclose personally identifiable information. OpenAI does not associate the data with specific individuals or develop user profiles.

Encryption Data transferred to and from OpenAI’s services undergoes encryption, ensuring secure communication.

Transparency OpenAI practises transparency regarding its data usage, allowing individuals to consult its privacy policy and terms of service for a thorough understanding of these.

7. Academic Integrity

CPUT’s Plagiarism Policy serves as the focal point for upholding academic integrity within the university. It demonstrates the university’s commitment to establishing a culture of scholarly academic conduct. The policy is aimed at ensuring that both academic staff and students are dedicated to producing those academic outputs that advance their respective fields of study. By alerting the CPUT community to the potential risks of using generative AI to plagiarism, this guide, informed by the CPUT Plagiarism Policy, aims to inspire a conscious effort among its staff and students to avoid academic misconduct. Ensuring the integrity of one’s work has grown increasingly difficult, while simultaneously becoming more critical in the age of LLMs. The distinction between original content and computer-generated material can occasionally become blurred when advanced technologies, such as ChatGPT, Google Gemini, or Microsoft BING become increasingly common. Thus, it is crucial for staff and students to learn to use these resources appropriately and ethically. Academic integrity may be under threat if the Plagiarism Policy, through the lens of generative AI, is not followed, or if generative AI is used carelessly without adequate and appropriate citation and understanding. Furthermore, while it is accepted that research frequently calls for expanding upon the ideas and contributions of others, CPUT’s Plagiarism Policy emphasises the need to harmoniously combine external knowledge with one’s original thoughts and analyses. By doing so, staff and students are not just paying homage to the contributions of their predecessors but are also ensuring the production and dissemination of authentic, original academic work.

The CPUT Plagiarism Policy defines plagiarism as the representation of another person’s ideas, research, expressions, computer code, design artefacts, or work as one’s own. Examples of plagiarism interpreted in the context of generative AI include, but are not limited to:

The CPUT plagiarism policy defines plagiarism as the representation of another person’s ideas, research, expressions, computer code, design artefacts, or work as one’s own. Examples of plagiarism interpreted in the context of generative AI include, but are not limited to:

Examples of plagiarism as listed in the CPUT plagiarism policy

• Copying directly from print or electronic sources into one’s own work.

• Imitating existing designs in one’s own work.

• Copying another student’s assignment or part thereof

• Overuse of sources; disguising the act of verbatim copying by substitution of wording

• Paraphrasing without citation

• Self-plagiarism (repeating own previously published research without acknowledgement).

Examples of plagiarism as listed in the CPUT plagiarism policy interpreted in the context of generative AI.

• Using generative AI models to produce content and submitting them as one’s own without citation. Using generative AI to scrape and reproduce content verbatim from databases, e-books, online journals, or websites without proper acknowledgment.

• Using generative AI-driven design software to reproduce or slightly modify existing designs, graphics, or art without significant original input, or without giving credit to the initial design source.

• Using generative AI tools to scan and replicate other students’ assignments or dissertation, even if modified slightly with generative AI assistance, and presenting such assignments as original work.

• Leveraging generative AI to paraphrase or reword content excessively from multiple sources without proper citation. This could be using tools that automatically swap words or reorder sentences to make content seem distinct, even if the core content remains unchanged.

• Employing generative AI-driven paraphrasing tools to alter the structure of a sentence or change words but maintaining the original meaning, and not citing the original source.

• Republishing one’s own content or research in multiple platforms or journals without indicating that it has been previously published or presented. This could also mean using generative AI tools to slightly modify one’s previous works and presenting them as new findings or creations.

The above examples of plagiarism from the CPUT plagiarism policy have now been interpreted in the context of generative AI. In the 21st century, AI tools can produce, replicate, or modify content, all of which pose challenges to academic integrity. These tools can be used to scrape content, imitate designs, replicate assignments, paraphrase without citation, and even modify one’s own previously published work. The development and increasing use of generative AI make it crucial for academic institutions to recognise these new dimensions of, and potential opportunities for, plagiarism and for them to adapt their guidelines and policies to maintain the sanctity of original work (OpenAI, 2022).

Although the CPUT Plagiarism Policy serves as the first means of ensuring academic integrity within the institution, the following sections may further strengthen and ensure academic integrity in the areas of assessment at CPUT.

7.1 Referencing Generative AI

When individuals seek to ensure accuracy in referencing materials within their academic or professional writing, adhering to the correct citation standards is paramount. In instances of uncertainty regarding the proper referencing of sources, a practical solution involves directly prompting a generative AI, such as OpenAI’s ChatGPT, for assistance on the subject. By posing specific inquiries related to citation formats for various types of sources, individuals can obtain customized guidance that complies with established referencing styles.

The screenshot below demonstrates how guidance on how to reference generative AI can be obtained

7.2 Assessment Design using Bloom’s Taxonomy

Focusing on the higher-order thinking skills in Bloom’s Taxonomy, skills that draw on the unique attributes of humans, can assist in the designing of the kinds of assessments which can mitigate unethical conduct. The table below illustrates the unique attributes and capabilities of humans and generative AI. The accompanying recommendations may help to ensure a comprehensive evaluation of student understanding and skills in the context of generative AI as well as helping to ensure ethical behaviour.

Blooms Taxonomy Human Generative AI Approach to mitigate unethical conduct

Remember

Understand

Can recall information in situations where technology is not readily accessible.

Apply

Analyse

Can contextualise answers applying emotional, moral, or ethical considerations.

Can operate, implement, conduct, execute, experiment, and test in the real world, applying their creativity and imagination to idea and solution development

Possess the unique capability to formulate original solutions, incorporating human judgement, and collaborating spontaneously.

Is adept at recalling mainly factual information, listing possible answers, defining terms, and constructing a basic chronology.

Is skilled at describing concepts in different words, recognising related examples, and translating.

Is capable of making use of a process, model, or method to illustrate how to solve a quantitative inquiry.

Design open-book exams or allow specific resources. Since generative AI can recall facts easily and instantly, design an assignment which tests the application or understanding of those facts rather than testing mere recollection. For instance, ask for unique experiences or personal interpretations related to a topic.

Ask students to explain topics in their own words, provide personal examples, or discuss what they perceive to be the moral implications of certain facts or theories.

Use scenario-based questions that require students to apply concepts in real-world situations. Encourage creative problem solving. For instance, “How would you apply theory X in situation Y?”

Can suggest a range of alternatives, enumerate potential drawbacks and advantages, and describe successful real-world cases.

Encourage students to design, plan, or propose solutions to openended problems. Group projects or collaborative assignments can also be effective, especially if they involve peer evaluations.

Evaluate Create Can engage in metacognitive reflection and holistically appraise the ethical consequences of alternative courses of action.

Possess the unique capability to formulate original solutions, incorporating human judgement, and collaborating spontaneously.

Can identify the pros and cons of various courses of action and is proficient at developing rubrics.

Ask students to evaluate or critique theories, methods, or case studies. For example, “Discuss what you see as the pros and cons of method X in context Y.”

Can suggest a range of alternatives, enumerate potential drawbacks and advantages, and describe successful real-world cases.

Encourage students to design, plan, or propose solutions to openended problems. Group projects or collaborative assignments can also be effective, especially if they involve peer evaluations.

7.3 Traditional Assessment Integrity

This list outlines traditional assessment types aimed at ensuring academic integrity and comprehensive student evaluation in the context of generative AI’s capabilities and limitations.

Assessment

On-campus Invigilated Assessments

Open-Book Exams

Make judicious use of supervised tests or exams on campus.

Develop assessments that prioritise comprehension rather than rote memorisation.

7.4 Online Assessment Integrity

This list outlines online assessment types aimed at ensuring academic integrity and comprehensive student evaluation in the context of generative AI’s capabilities and limitations.

Assessment tools and strategies

Turnitin, and Respondus

Use well-known tools for detecting plagiarism in LockDown Browser assignments and ensuring honesty during online exams.

Randomised Questioning

Platforms that randomise exam questions to deter collaboration.

7.5 Feedback and Review Mechanisms

The recommendations below are aimed at fostering a culture of accountability and continuous improvement, thereby strengthening academic integrity by making it more challenging for breaches to go ignored.

Feedback and monitoring

Peer Review

Feedback Mechanisms

Promote peer evaluation among students to foster a sense of group accountability.

Consistent, and constructive feedback motivates genuine effort and improvement.

Monitor the Process Especially in smaller classes, tracking drafts and reflections ensures original student work.

7.6 Honour Codes and Educational Workshops

The following is a list of Initiatives to promote academic integrity.

Promotion and awareness strategies

Honour Codes

Educational Workshops

Example

Establish a culture of integrity where students pledge against academic dishonesty.

Create awareness through workshops on academic integrity and plagiarism. The Fundani Writing Centre has numerous workshops that can assist with developing this.

7.7 Additional Recommendations

The following are additional recommendations to ensure academic integrity.

Resources and assessment methods Example

Resource Availability

Oral Examinations

Offering writing centres and tutoring can reduce the need for dishonest conduct.

The use of oral examination may also mitigate dishonest conduct.

8. Planning and Preparation for Teaching

Generative AI has emerged as a transformative tool that can significantly aid teaching in various aspects of planning and preparation. Judicious leveraging of AI technologies can streamline the intricate process of designing curricula, generating teaching materials, and ensuring quality educational delivery (OpenAI, 2022). Below are some recommendations for the use of generative AI to assist in planning and preparation.

Create lesson plans

Draft structured plans that detail the flow and content of a lesson.

Write lesson outcomes

Develop a worksheet

Create PowerPoint

Define what students should understand or achieve by the end of a lesson. Create custom questions or tasks related to a specific topic.

Ask generative AI to “Draft a lesson plan for a 45-minute session on Shakespeare’s ‘Romeo and Juliet’ for first year 10 students.” The AI’s plan might include an introduction, group activities, and more.

Give generative AI an outline on “World War I” and ask, “What are the expected learning outcomes after this lesson?” The generative AI might suggest outcomes like “Understand the complexity of the causes of World War I.”

Prompt generative AI with, “Generate ten multiple-choice questions about the water cycle for a first-year class.” Use, and possibly adapt, the AI’s generated questions for a worksheet.

Identify the ‘big ideas’ of a course, unit, or lesson

Transform textual content into structured points suitable for visual presentation.

Extract key concepts from lengthier materials to focus on essential ideas.

Give generative AI lecture notes on “Photosynthesis” and instruct, “Convert these notes into bullet points for a PowerPoint presentation.” Use the produced list for slide creation.

Input an article about “The Renaissance” into generative AI and prompt, “Summarise the key points of the Renaissance.” Use the generative AI’s response to pinpoint core concepts for your lesson.

Evaluate the output for correctness and completeness

Generate benchmark answers or solutions to streamline grading.

For an essay on climate change, feed generative AI the topic and request a model essay. This can then serve as a comparison point for student submissions.

Consider the student’s perspective

Anticipate areas of confusion or potential questions by simulating student responses.

Before teaching “quantum physics”, use generative AI to simulate student queries. Input, “Explain quantum physics like I’m a first-year student.” The AI’s answers can guide and/or suggest your teaching approach.

9. The Use of Generative AI in Learning and Teaching

As with the previous sections, the tables below were developed using generative AI. Furthermore, invaluable input was contributed by colleagues from CPUT, reflecting a collaborative effort that combines academic insights with cutting-edge technological advancements. This list recognises the dynamic and evolving landscape of educational technology, aiming to equip educators and learners with tools and methods that are relevant to today’s digital age. Importantly, as the realm of educational technology continues to advance, this table remains adaptable and poised for updates and refinements in response to new developments.

9.1

Information Processing

These recommendations focus on refining and reworking raw data or extensive information. Tasks here involve condensing, summarising, and rephrasing to make content clearer.

Use of generative AI Description

Sum up information Condense extensive information

Example of prompt to generative AI

Summarise the main themes of ‘The into a concise summary. Great Gatsby’ in 200 words.

Briefly explain Keynesian economics in 250 words.

Rephrase tricky ideas Reformulate difficult concepts Word thoughts on post-modernism. or thoughts into clear language. Rephrase the main points of an argument.

9.2 Academic Support

The subsequent list of recommendations offers a comprehensive overview of the multifaceted roles that AI can play in aiding both teaching and learning, ranging from providing scaffolding support and solving mathematical problems, to assisting with test preparation and writing reviews.

Use of generative AI Description

Provide scaffolding support

Solve math problems using steps

Explain big concepts succinctly.

Plan out writing pieces

Review writing Check my language quality

Prepare for tests

Explain code or formulas

Assist with assignments

Identify prior knowledge required to understand a topic

Break complex processes into smaller, more manageable steps.

Calculate mathematical problems and show the methodology.

Offer brief, accessible explanations of complex subjects.

Provide structure or an outline for writing assignments.

Assess and provide constructive feedback on written content.

Evaluate text for clarity, consistency, and grammatical correctness.

Create materials to aid in test preparation.

Describe and illustrate coding concepts or mathematical formulas.

Offer guidance and solutions for academic assignments.

Determine the extent of familiarity with the foundational concepts needed before exploring advanced subjects.

Example of prompt to generative AI

Simplify the process of photosynthesis. Guide through stoichiometry in chemistry.

Work out 3x - 7 = 11 with steps

Find the derivative of f(x) = x^3 + 4x^2 - 5x and explain how to do this.

Give a short explanation of existentialism. Give an overview of quantum mechanics.

Topics to cover about renewable energy. Essay outline on city growth challenges.

Review paragraph on gender roles.

Feedback on the content of an essay about climate change.

Review the research abstract for clarity. Ensure the use of consistent tenses in an essay.

Flashcards on the French Revolution. Multiple-choice questions on integration. Explanation of a ‘while’ loop in C++. How VLOOKUP works in Excel.

Help with physics homework on Newton’s laws. Guidance on a calculus problem about integrals.

Foundational concepts for advanced genetics. Background knowledge before studying the French Revolution.

9.3

Language and Writing

The following recommendations provide an overview of how generative AI can serve as a tool in educational settings, focusing on its capabilities to generate ideas for writing topics, assist with language and grammar, and even create multilingual glossaries for specific subjects.

Use of generative AI Description

Give ideas for writing topics

Help with language and grammar

Create a glossary of terms for a topic in different languages

Suggest creative themes or topics for writing.

Assist in correct language usage, language registers, and grammatical structures.

List and translate key terms related to a subject in multiple languages.

Example of prompt to generative AI

Story idea for a post-apocalyptic world. Creative writing prompt about time travel.

Sentence with the word ‘albeit’.

Proper use of ‘whom’.

Glossary of anatomy terms in English and Xhosa (isiXhosa).

Physics terms related to ‘Motion’ in English and French.

9.4 Personalised Learning

The following are recommendations for how generative AI can open up avenues for tailoring personalised learning paths, crafting engaging visual aids, simulating experiments or phenomena, and conducting insightful analyses of student performance with valuable feedback.

Use of generative AI Description

Personalised Learning Paths

Generate Visual Aids

Simulate Experiments or Phenomena

Analyse Student Performance and Give Feedback

Tailor learning trajectories based on individual needs and progress.

Produce graphical representations to explain concepts.

Digitally reproduce scientific events or phenomena for understanding.

Evaluate learners’ progress and provide constructive, developmental feedback.

Example of prompt to generative AI

Algebra topics following on from, and based on, current understanding. Exercises for improving reading comprehension.

Visual representation of the water cycle. Flowchart for photosynthesis.

Simulation of pendulum motion.

Demonstration of volcanic eruptions.

Progress assessment in geometry.

Feedback on academic performance for the semester.

9.5 Resource Recommendations

The following is a list of recommendations to guide students to additional materials using generative AI. Generative AI can suggest books, documentaries, or other resources. It can aid in expansion of a student’s knowledge base outside of the primary course content.

Use of generative AI Description

Provide Recommendations for Additional Resources

9.6

Suggest supplementary materials for deeper exploration of topics.

Example of prompt to AI

• Book recommendations on African Architecture and Design. Documentary suggestions on climate change.

Collaborative Learning

The following is a list of recommendations for ways in which generative AI can support group interactions and peer-to-peer engagements. These recommendations focus on activities that foster teamwork, discussion and shared learning experiences.

Use of generative AI Description

Facilitate Group Collaborations and Peer Learning

Gamified Learning and Quizzes

Enable collaborative learning experiences and peer-to-peer interactions.

Design engaging and interactive learning exercises and assessments.

Example of prompt to generative AI

Set up space for group projects.

Connect peers for learning sessions.

Quiz on cellular biology.

Gamified exercise.

Further to the above recommendations, in a recent article by Mike Sharples (2023), “Towards Social Generative AI for Education: Theory, Practice and Ethics”, generative AI is described as a component in an education system in which humans and AI act as language agents. Sharples (2023) argues for conversation as a fundamental process in learning. He argues that, through conversation, we can reflect on existing understandings of knowledge and question underlying assumptions as we come to a consensus. Through a conversational learning system, the potential exists to continually explore differences, gain experience, and reach agreements (Sharples, 2023). Generative AI can enable these collaborative, ‘conversational’ learning experiences.

Sharples references six different roles of generative AI supporting cooperative and social learning. These roles can be realised in both a synchronous face-to-face lecture theatre modality and through an online modality. In the lecture theatre students and lecturers can together provide the prompts for the generative AI. In the online modality of a lecture stream, using platforms such as Microsoft Teams and Blackboard Collaborate, students could be encouraged to provide prompts either via a microphone or chat functionality. As outlined in the earlier sections, with the use of generative AI, it is essential to provide a clear rationale and instructions, ensure student understanding, promote critical thinking, emphasise the importance and ethicality of verifying information, and encourage open and rigorous debate.

The roles and functions of generative AI in such processes include a Possibility Engine, Socratic Opponent, Collaboration Coach, Co-Designer, Exploratorium, and Storyteller.

Role of generative AI Description

Possibility Engine

Socratic Opponent

A student may explore an open question, prompting generative AI to generate multiple responses.

Generative AI can be used as an opponent in an argument.

Example of prompts

Students write prompts in generative AI and submit each prompt multiple times to examine alternative responses.

Collaboration Coach

Co-Designer

Exploratorium

Storyteller

Generative AI helps groups to research and solve problems together.

Generative AI assists throughout the design process.

Generative AI provides tools to play with, explore, and interpret data.

Generative AI creates stories that include diverse views, abilities and experiences.

Students enter prompts into generative AI to converse or debate.

Teachers can ask students to use generative AI to prepare for discussions.

Working in groups, students use generative AI to discover information to complete assignments.

Students ask generative AI for ideas about designing or updating a website, or focus on specific goals (e.g., how to make the website more accessible).

Students use generative AI to explore different ways to visualise and explain a large database, such as census data.

Students take it in turn to ask generative AI to continue a story, prompting it to include a diversity of characters.

10. The Use of Generative A I in Research

As outlined in earlier sections of this guide, the utilisation of generative AI for teaching and learning often intersects with its potential in research. However, the following table offers examples of how generative AI can distinctly support research activities.

Research Activity

Literature Review

Example of prompt Description Research Application

“Summarise key themes from the latest articles on [subject area].”

Hypothesis

Generation Data Analysis

Drafting Research Proposals

“Suggest potential hypotheses based on the current understanding of [topic].”

“Analyse the dataset [details] and identify any significant patterns.”

“Draft an outline for a research proposal on [topic] including objectives and methodology.”

Experiment Design

Statistical Analysis Research Collaboration

Writing Research Papers

Peer Review Preparation

Enhancing Research Presentations

“Design a research experiment for [study objective] considering control and variables.”

“Perform a statistical analysis using [statistical method] on [data].”

“Identify potential collaborators with expertise in [field] and [specific interest].”

“Generate a draft for the discussion section of a paper based on [findings].”

“Create a checklist for peer-reviewing a manuscript on [topic].”

“Suggest a structure for a presentation on [research findings], including visual aids.”

Generative AI will provide condensed information from recent publications within a subject area.

Generative AI uses existing knowledge to propose plausible research questions or hypotheses.

Generative AI tools process and interpret complex datasets to find meaningful patterns or anomalies.

Generative AI helps in structuring a research proposal, suggesting a coherent flow of ideas.

Generative AI can suggest experimental setups that align with research goals.

Generative AI can carry out complex statistical computations and interpret the results.

Generative AI can suggest researchers and professionals as potential collaborators.

Generative AI can help in structuring the narrative of research findings.

Generative AI develops a comprehensive list of aspects to consider when reviewing a paper.

Generative AI helps in organizing content and suggests visualizations for clarity.

Helps in grasping the current state of research and identifying trends.

Assists in developing testable predictions to guide the research direction.

Aids in making sense of large amounts of data, which can be time-consuming manually.

Provides a starting point for proposal writing, which can be refined further.

Helps in creating robust experimental designs that can lead to valid results.

Saves time and improves accuracy in the data analysis phase of research.

Facilitates the formation of multidisciplinary teams for research projects.

Assists in articulating the implications and significance of research results.

Ensures thorough preparation for peer review, contributing to research quality.

Enhances the delivery and impact of research findings in presentations.

11. Ai Artistry and V ideo G eneration

For those interested in the intersection of language and visual representation, DALL-E is a groundbreaking technology you should explore. Developed by OpenAI, DALL-E is a state-of-the-art image generation model that combines the capabilities of GPT-3 with a VQ-VAE-2 generative model. This allows it to produce highly detailed and imaginative images based on textual descriptions. Whether you’re looking to better understand the frontiers of machine learning or are considering its implications in various academic disciplines, DALL-E offers a fascinating look into the future of AI-generated art and design.

Figure 6: Discover a great guide to mastering DALL-E: from aesthetics and art history to prompt engineering and image editing. Unlock endless creative possibilities! Access the guide here: https://dallery.gallery/the-dalle-2-prompt-book/

Developments in the area of LLMs have been rapid, and these can now generate videos from natural language prompts as well. OpenAI’s Sora offers this capability, as well as Bing Chat Create and Pika (https://pika.art)

12.Conclusion (Written by generative AI)

The integration of generative AI into the academic environment of CPUT heralds a new era of educational innovation. The technological advancements represented by tools like generative AI have the potential to significantly transform and enrich pedagogical approaches, curriculum design, and academic integrity. While generative AI presents unprecedented opportunities to enrich learning and teaching, it also brings forth ethical and integrity challenges that require astute navigation. The university’s proactive stance, as evidenced by this comprehensive guide, demonstrates a commitment to harnessing the power of generative AI while safeguarding academic integrity. generative AI, when used responsibly and ethically, can serve as a tool to address some of these challenges, offering both educators and students a means to navigate the evolving landscape of knowledge dissemination. In conclusion, as CPUT embarks on this journey of weaving generative AI into its academic fabric, it does so standing at the precipitous edge of a transformative educational experience. With the appropriate and imaginative guidelines, continuous discourse, and ethical considerations, the university is well prepared to lead in this AI-augmented future, and to ensure that its students and academic staff are fully equipped to navigate the complexities of the modern world.

13. Additional Resources

This list of resources serves as an additional guide to explore the intersection of AI and education. Covering a range of topics from foundational AI concepts for educators to specific applications in the classroom, these resources have been curated to provide a comprehensive understanding of the role AI can play in transforming educational practices. The list includes research and initiatives from both South African universities and global institutions to offer a well-rounded perspective on the subject.

Stanford’s Artificial Intelligence in Education Report: This is an extensive report covering how AI can be integrated into K-12 education. https://hai.stanford.edu/research/education/

University of Cape Town’s AI and Learning Research: Papers and articles focused on the use of AI in South African educational settings. URL: https://www.uct.ac.za/research/ai-education

AI4K12 Guidelines: A framework for educators in K-12 on how to effectively implement AI in classrooms. URL: https://ai4k12.org/

Coursera’s AI for Everyone: While not directly related to education, this course teaches foundational AI concepts that educators could implement. URL: https://www.coursera.org/learn/ai-for-everyone

IBM’s Teacher Advisor: An AI tool designed to assist teachers in creating lesson plans. URL: https://teacheradvisor.org/

WITS AI in Education: Research and initiatives from the University of the Witwatersrand in Johannesburg focusing on AI in education. URL: https://www.wits.ac.za/research-ai/

AI in Higher Education by Educause Review: An article focusing on how AI can benefit tertiary education. URL: https://er.educause.edu/articles/2018/10/artificial-intelligence-in-higher-educationapplications-promises-and-perils

MIT’s Open Learning: AI-based projects aimed at transforming teaching and learning. URL: https:// openlearning.mit.edu/

Google’s Machine Learning for Kids: A tool that lets children and educators experiment with machine learning. URL: https://machinelearningforkids.co.uk/

Udacity’s AI for Classroom: Online courses designed for teachers interested in AI. URL: https://www.udacity.com/course/ai-for-classrooms--ud0419

Harvard’s Data Wise Project: A guide to help educators use data, including AI-generated analytics. URL: https://datawise.gse.harvard.edu/

European Commission’s AI in Education: Provides information about AI in educational practices in Europe. URL: https://digital-strategy.ec.europa.eu/en/policies/ai-education

University of Pretoria’s Research in AI and Learning: Includes several research papers and pilot projects. URL: https://www.up.ac.za/research-innovation

Learning LandsCAPE AI for Research symposium: https://www.dropbox.com/scl/fi/ ydgjhv9egao3jj2v6s366/AI-Research-tools-Tools.pdf?rlkey=3u6fityrmf06p3x6almvsx83d&dl=0

Microsoft’s AI in Education: A set of tools and solutions specifically designed for educational applications. URL: https://education.microsoft.com/en-us/resource/520e08e6

Squirrel AI Learning: Offers adaptive learning tools for students, providing personalised education. URL: https://www.squirrelai.com/en/index.html

Oxford’s AI Education Research: Various papers and articles on AI in educational settings. URL: https://www.oii.ox.ac.uk/research/ai/

DeepMind’s AI for Education White Paper: Discusses the potential benefits and ethical considerations of using AI in education. URL: https://deepmind.com/research/publications/ai-education

Australian Government’s AI Ethics in Education: A report discussing the ethical dimensions of AI in education. URL: https://www.data61.csiro.au/en/Who-we-are/Our-programs/AI-and-machinelearning/AI-Ethics

Querium’s AI-based Tutor: An AI system that acts like a tutor for STEM subjects. URL: https://www. querium.com/

Elementary Literacy through AI: Developed by Carnegie Mellon, this tool aims to improve literacy in elementary schools using AI. URL: https://www.cmu.edu/news/stories/archives/2019/february/ailiteracy.html

14. References

Dallery Gallery. (2023, October 26). The DALL-E 2 Prompt Book. https://dallery.gallery/the-dalle-2prompt-book/

OpenAI. (2023). ChatGPT. OpenAI. https://www.openai.com/

Paul, R., & Elder, L. (2006). Critical thinking: The nature of critical and creative thought. Journal of Developmental Education, 30(2), 34.

Sharples, M. (2023). Towards social generative AI for education: theory, practices, and ethics. arXiv:2306.10063.

West, D, M. (2018). The Future of Work: Robots, AI, and Automation: Brookings Institutional Press.