Issuu

NEWS

CULTURE

FEATURE

OPINION

The physics team successfully defended their first place title in the United State Invitational Young Physics Tournament.

Read about what made the grade and what failed in this issue’s edition of the Pop Culture Report Card.

TikTok is facing a proposed ban from the Biden administration amid escalating U.S.-China tensions.

Gabe A. ’24 believes that we should re-examine the stigma around obtaining outside care and assistance for elders.

READ MORE ON PAGE 16

2/16/2023

THE

NUEVA CURRENT THE NUEVA SCHOOL, SAN MATEO, CA 94403

INSTAGRAM & TWITTER: @THENUEVACURRENT

ISSUE FOUR

VOLUME 6

IN DIALOGUE WITH AI Students and faculty engage in conversations about the possibilities and ramifications of ChatGPT

STORY ART

Emma Z. & Aaron H. Anwen C.

I. WHAT IS CHATGPT? The prompt was the same as it had been every year that she taught it to her English 11 students: “Craft an argumentative essay on Frederick Douglass’ biography.” This time though, she was feeding it to her computer. With curiosity and apprehension, Upper School English Teacher Amber Carpenter hit the “return” key on her keyboard. Her cursor blinked once, then twice, then began to move. For years, Carpenter had taught high school and college students to craft a well-written essay—to develop a nuanced argument, structure it tactfully, and supplement it with the proper writing devices. So as she watched the ability of the newly released chatbot ChatGPT unfold across her screen— an insentient machine learning algorithm generating a comprehensive essay in mere seconds—she felt simultaneous awe and trepidation. Carpenter was not alone in her shock. ChatGPT was developed by OpenAI, a San Francisco-based research company co-founded by Elon Musk in 2018 and backed by Microsoft, and became an immediate viral sensation upon its release on Nov. 30, 2022. Trained on a whopping 570 GB of text data from major online databases such as Wikipedia (166 million words comprise one GB on average), the know-it-all chatbot took the internet by storm with its impressive linguistic abilities and apparent creative capacity. If prompted, ChatGPT can generate analytical essays, explain complex mathematical concepts in simple terms, and “write a biblical verse in the style of the King

James bible explaining how to remove a peanut butter sandwich from a VCR.” By the fifth day of its release, ChatGPT had amassed more than a million registered users—a feat that technology giants such as Spotify, Facebook, and Instagram took months to achieve. Less than two months later, the chatbot hit an estimated 100 million monthly active users. And now, ChatGPT is the fastest growing consumer internet application in history, according to a UBS study. Up until now, AI bots available to the public (such as Apple Inc’s Siri, Amazon’s Alexa, Microsoft’s Cortana, and Google’s Google Assistant) have been notoriously sub-par. In comparing them to today’s ChatGPT, upper school math teacher Ted Theodosopoulos—who has been involved in work with computational linguistics for several decades—recounts that they were “clunky” and “unnatural.” “I understand how [ChatGPT] works, so I wasn’t surprised by what it outputted,” said Theodosopoulos, mostly unenthused about his first and only venture into the AI. “But at the same time, I’m quite impressed with how far they’ve come; ChatGPT feels much more fluid than earlier chatbots.” One reason for ChatGPT’s fluidity is that, unlike many previous chatbots, it can recall previous dialogue. In many ways a casual conversation with ChatGPT, ranging in topic from the weather to the meaning of life, can feel real and personalized. Want to know the secret to good guacamole? Ask ChatGPT. “Use ripe avocados,” “mash them by hand,” and add cilantro, red onions, and ripe tomatoes just before serving “to keep their flavor and texture intact,” are just a few tidbits of wisdom that the chatbot imparts. “What if I don’t like tomatoes?” “If you don't like tomatoes, you can easily

leave them out of your guacamole recipe,” ChatGPT responds, before kindly suggesting jalapeños or bell peppers as a replacement. But like all chatbots, ChatGPT has its limitations. Because it is trained off of data from the nebula of the internet to string together text sequences, it’s not only unable to create anything genuinely new, but also dangerously prone to incorrect responses. OpenAI acknowledges these shortcomings in the software’s “Limitations” section alongside its capabilities and example prompts on the main chat window: “May occasionally generate incorrect information,” “May occasionally produce harmful instructions or biased content,” and “Limited knowledge of world and events after 2021.” Yet, beyond ChatGPT’s potential factual inaccuracies, teachers such as Carpenter fear that ChatGPT poses a threat to academic integrity and student motivation in classes. If ChatGPT was powerful enough to generate unique, articulate text samples, what was stopping students from using ChatGPT to complete their essays? And if ChatGPT could get the job done, did students even need to learn how to write anymore? Soon after the AI’s release, opinion essays on ChatGPT written by teachers began to populate news feeds. One article in The Atlantic is titled “The End of High School English,” another, “The College Essay is Dead.” Popular news sources such as The New York Times have since published rebuttals to both articles, but the fear of writing-class obscurity remains. On Dec.14, Furman University assistant philosophy professor Darren Hick posted on his facebook page about using OpenAI’s AI text classifier to implicate one of his students, who he suspected had used ChatGPT to write a take-home test. The classifier had been “99.9

percent” sure that the work was fake. “Today, I turned in the first plagiarist I’ve caught using A.I. software to write her work,” he wrote. “Administrations are going to have to develop standards for dealing with these kinds of cases, and they’re going to have to do it FAST… This is too new. But it’s going to catch on… Expect a flood, people, not a trickle.” Several academic communities have already jumped into action to address these concerns. In early January, New York City Public schools—the largest school district in the U.S.—banned ChatGPT across all district devices and networks. Other large city districts such as Seattle and Los Angeles have since followed in their footsteps. At the same time, some educators see benefits in working alongside emerging technology like the chatbot, and have published their own opinions. On January 12, Kevin Roose wrote an Op-Ed for The New York Times headlined “Don’t Ban ChatGPT in Schools. Teach With It.” “After talking with dozens of educators over the past few weeks, I’ve come around to the view that banning ChatGPT from the classroom is the wrong move,” Roose argues. “Instead, I believe schools should thoughtfully embrace ChatGPT as a teaching aid—one that could unlock student creativity, offer personalized tutoring, and better prepare students to work alongside A.I. systems as adults.” Amidst the heyday of differing opinions, the overarching question remains the same, and appears to be one that the seemingly all-knowing chatbot can’t answer: What does ChatGPT mean for the future of education? Does that answer change at a “gifted” school like Nueva? CONTINUED ON PAGE 12