Why Even Basic A.I. Use Is So Bad for Students
Author: Anastasia Berg
Berg, Anastasia. “Why Even Basic A.I. Use Is So Bad for Students.” The New York Times, 29 Oct. 2025, www.nytimes.com/2025/10/29/opinion/ai-students-thinking-school-reading.html.
0 General Document comments
0 Sentence and Paragraph comments
0 Image and Video comments
Last spring, it became clear to me that over half the students in my large general education lecture course had used artificial intelligence tools, contrary to my explicit policy, to write their final take-home exams. (Ironically, the course was titled Contemporary Moral Problems: The Value of Human Life.) I had asked them about some very recent work in philosophy, parts of which happened to share titles with entirely different ideas in medieval theology. You can guess which topics the students ended up “writing” about.
My situation was hardly unique — rampant A.I. cheating has been reported all over the country. But I felt a dread I struggled to express until a colleague articulated the problem in stark terms: “Our students are about to turn subcognitive,” she said. That was it. At stake are not just specialized academic skills or refined habits of mind but also the most basic form of cognitive fluency. To leave our students to their own devices — which is to say, to the devices of A.I. companies — is to deprive them of indispensable opportunities to develop their linguistic mastery, and with it their most elementary powers of thought. This means they will lack the means to understand the world they live in or navigate it effectively.
A.I. is hardly the first technology to threaten our cognitive competence. Long before ChatGPT, the smartphone and the calculator, Plato warned against writing itself. Literate human beings, he foresaw, would “not use their memories.” He was not entirely wrong. But few of us would consider this a bad bargain. The written word is, after all, the condition for the survival of these very same Platonic dialogues across two millenniums. Great gifts have often come at great cost. The question is always: Are they worth it?
As students’ A.I. use has proliferated, many of its critics focused on intellectual gifts. “A.I. undermines the human value of attention,” the poet Meghan O’Rourke wrote in a guest essay for Times Opinion, “and the individuality that flows from that.” Other endangered powers: “unique human expression,” “the slow deliberation of critical thinking” and the “ability to write original and interesting sentences.” As a humanities professor, all these concerns resonate with me.
Yet I have come to see that something far more fundamental is being put at risk. Developing our linguistic capacities — to master diverse concepts, to follow an intricate argument, to form judgments, to communicate those to others — is the development of our capacity to think.
For us human beings, using language is not a skill like any other — it is the way we do almost anything at all. Philosophers have disputed whether beings could exist that could think despite lacking language, but it is clear that humans cannot do so. We grasp the very contours of our world in and through language. But we are not born with a language. We have to acquire and develop our linguistic capacities through immersive practice with other human beings. For hundreds of years, in advanced societies this has meant cultivating an intimate familiarity with human writing.
Many people invoke a distinction between illicit uses of A.I. (such as the composition of entire drafts) and innocent auxiliary functions — outlining, for instance. But it is these seemingly benign functions that are the most pernicious for developing minds. Take the summary: Letting A.I. take over this rote task seems like a harmless shortcut. Sure, students who read only A.I. summaries will be subjected to predictable analysis and homogeneous prose, but they could save time and energy. In truth, the ability to determine what is being argued for and how is not dispensable. No aspect of cognitive understanding is perfunctory.
Without the chance to develop these capacities, young people would not be able to understand a news report, medical documents and consent forms or the merit of an argument (including this one). A depleted conceptual reservoir would render our lives crude and our experience of the world undifferentiated and coarse. Worst of all, cognitive degradation threatens our claim to self-rule: It is far from obvious that the denizens of the subcognitive society would be fit to participate in the democratic processes that determine how we structure our societies and lives.
Granting its potential downsides, many educators insist that widespread A.I. use in education is inevitable. A.I. enthusiasts such as Princeton’s D. Graham Burnett claim that literacy is a historical aberration, no longer fit for this world. Outside a few elite institutions, Dr. Burnett argues, there will soon be no point in asking students to read books. Teachers should instead ask students to “do stuff” with short texts: “Sing them. Memorize them. Cut them up into little pieces and stick them on the walls.” In other words, we ought to welcome a return to a society where functional literacy is reserved for the privileged few.
This seems like a pretty easy argument to attack. It turns AI use by students into an either-or. But there are some nuances in between. For example, students can still choose to use it or not at different times and for different purposes. She doesn’t address some of the stronger points from AI in education advocates.
I bristle at Dr. Burnett’s blithe consignment of the majority of American college students to an education that seems better fit for kindergartners. And his claims are based upon a false premise. Many students can still read and engage with long texts — and want to. Just look at the half of my class who did that rather than use A.I. tools. And none of my students — whether they read the challenging texts I assigned them or not — would have been better served by using small bits of our essays for wall decoration instead.
Higher education aims to create cognitively mature adults, which in turn requires us to ensure students learn to read, think and write all on their own. It is easier than we think: Creating tech-free spaces and incentivizing students to spend time in them requires no new resources. All it takes is will. Many of our students still have it. Do their teachers?
General Document Comments 0
Here’s my quick SIFT snapshot: This essay comes from a thoughtful, first-person perspective and includes references to other sources, but the strong emotional tone and lack of balance suggest I should look more closely at the author’s background, the context of the links, and whether expert consensus aligns with these claims before accepting anything at face value.
Would you like to start by pausing and reflecting on your first reactions to the opening of this essay, especially the line: “over half the students in my large general education lecture course had used artificial intelligence tools…”? What stands out to you or triggers a strong response? ✅
How did you feel or what did you assume when you read the author’s statement that “over half the students in my large general education lecture course had used artificial intelligence tools”—did this surprise you, and why do you think the author used this example at the beginning? ✅
Now let’s check out the author’s background together—can you find the name of the author of this essay, look them up on Wikipedia or their university page, and tell me what you discover about their expertise, possible biases, or publication history? Share the author’s name, title, and a relevant link. ✅
You mentioned Anastasia Berg, so what did you find out when you searched for her on Wikipedia or her university profile—what does her background as an Assistant Professor of Philosophy and Editor of The Point magazine tell you about her possible perspective, and do you notice any patterns or potential biases in the types of topics she’s written or published on before? (Please give me her title and a link to a relevant profile or publication.) ✅
Based on what you found—Anastasia Berg is an Assistant Professor of Philosophy and an editor at The Point but doesn’t have a record of publishing much about technology or AI—how might her expertise in philosophy and frequent writing about family or ethical issues shape the perspective or tone of her essay on AI in education, and do you see any connections or possible gaps? (Please share the link you referenced for her bio or publication list.) ✅
Now let’s try to “find better coverage”: Can you use AllSides Search or another news aggregator to find at least two other recent articles about AI and student cheating in college—do these sources agree with Berg’s concerns, or do they provide different viewpoints or evidence? (Please share each article’s title, author, and link.) ✅
Here’s something from Chalkbeat that feels real and nuanced.
https://www.chalkbeat.org/2025/11/04/three-theories-on-ai-in-schools-about-cheating-teaching-and-tutoring/ 3 ways AI will (and won’t) change schools
By Matt Barnum | November 4, 2025, 10:49am EST’
THIS! This is an article I want to learn more from:
https://socialinnovationsjournal.com/index.php/sij/article/view/9995/8125
Reimagining Writing Assignments in the Age of GenAI: A Practical Guide for EducatorsBy: Amy Stornaiuolo1, Robert Jean LeBlanc2
Great start! Now, looking at these articles—one from Chalkbeat and one from Social Innovations Journal—how do their tones, approaches, or solutions to AI and student cheating compare to those in Berg’s essay, and do these pieces suggest different conclusions about the risks or opportunities of AI in education? (Be sure to include the titles, authors, and links when you make your comparison.) ✅
After reading “3 ways AI will (and won’t) change schools” by Matt Barnum (link), and “Reimagining Writing Assignments in the Age of GenAI: A Practical Guide for Educators” by Amy Stornaiuolo and Robert Jean LeBlanc (link), how does their coverage of AI in schools—including concerns, responses, and tone—differ from or align with Berg’s more alarmed and philosophical warnings about “subcognitive” students? (Use specific comparisons.) ✅
Now let’s dig into the original claims and citations in the Berg essay—pick one claim (for example: “rampant A.I. cheating has been reported all over the country”) and see if you can trace it back to the original report or source it cites, then check if that original evidence matches how it’s used here (please share what you find and cite your source). ✅
There is significant evidence showing that AI-enabled cheating has become a widespread concern in educational institutions across the United States, but the claim of it being “rampant all over the country” should be interpreted with nuance and context. Multiple news outlets, educator reports, and research studies confirm notable increases in AI-related academic dishonesty, although the data varies by school type, region, and detection methods12345.
- Reported Incidents and Trends
- Major sources such as Axios and The Wall Street Journal cite teachers and administrators describing AI-linked cheating as rampant, with cases in both high schools and colleges forcing educators to regularly act as “AI detectors”12.
- Surveys indicate that between 6.4% and 24.1% of high school students admit to using AI tools or unauthorized devices to cheat, with the highest rates observed in charter schools4.
- AP News and other outlets state that over 60% of teachers witnessed an increase in students getting in trouble for using AI to complete assignments in the 2023-24 school year3.
- In some college surveys, about 40% of students admitted to using AI tools like ChatGPT for homework without permission2.
- High-profile cases have emerged, such as large classes being caught cheating with AI tools and even sending AI-generated apology emails to professors, stirring concern among educators67.
- Academic Context, False Positives, and Policy Response
- The reality of “rampant” depends on detection and reporting: many wrongful accusations occur because AI detection tools can mistakenly flag original work as AI-generated86.
- While confirmed breaches involving unauthorized AI use are high, many suspected cases are overturned when challenged, showing the difficulty of accurate enforcement8.
- Some universities project that cheating cases are undercounted because a majority of AI-generated work evades detection, with estimates that up to 94% of AI-authored submissions go unidentified4.
- Schools nationwide are adjusting policies and curricula to address the new challenges posed by generative AI. There remains no consistent national standard for policing AI cheating, leading to significant confusion and inconsistency in how cases are handled125.
- Summary Table: Reported AI Cheating Evidence
| Source/Survey Scope | Incidence Rate | Notes |
| -—————————————- | -—————————————————- | -————————————————————————- |
| High School (Nerdynav, 2025) | 6.4%-24.1% admit AI cheating4 | Highest rates in charter schools |
| College (WSJ/AP News) | ~40% AI use without permission2 | Growing national concern |
| Teacher Survey (CDT, 2024) | 63% report AI-related discipline3 | Sharp increase over previous year |
| UK Universities (Guardian) | 7,000 cases, triple prior year4 | Severe undercount suspected |
- Conclusion
Substantial evidence corroborates that AI cheating is a serious and pervasive issue, frequently described as “rampant” in media and educator statements, though the precise scale varies substantially by context12345. The label “reported all over the country” is largely accurate, but estimates can be inflated or inconsistent due to false positives, underreporting, and evolving policies86.
Sources
1 AI cheating surge pushes schools into chaos – Axios https://www.axios.com/2025/05/26/ai-chatgpt-cheating-college-teachers
2 AI’s Classroom Takeover: Cheating Crisis or Educational Revolution? https://www.webpronews.com/ais-classroom-takeover-cheating-crisis-or-educational-revolution/
3 New Data Reveal How Many Students Are Using AI to Cheat https://www.edweek.org/technology/new-data-reveal-how-many-students-are-using-ai-to-cheat/2024/04
4 ChatGPT Cheating Statistics (2025): Latest Facts on AI in … – Nerdynav https://nerdynav.com/chatgpt-cheating-statistics/
5 What counts as cheating with AI? Schools grapple with drawing the … https://www.latimes.com/california/story/2025-09-27/ai-students-cheating-tests-teachers-respond-lausd
6 Professors Aghast as Class Caught Cheating “Sincerely” Apologizes … https://futurism.com/artificial-intelligence/class-caught-cheating-apologizes
7 Their Professors Caught Them Cheating. They Used A.I. to Apologize. https://www.nytimes.com/2025/10/29/us/university-illinois-students-cheating-ai.html
8 University wrongly accuses students of using artificial intelligence to … https://www.abc.net.au/news/2025-10-09/artificial-intelligence-cheating-australian-catholic-university/105863524
9 AI Plagiarism Statistics 2025: Transforming Academic Integrity https://artsmart.ai/blog/ai-plagiarism-statistics/
10 AI Cheating Is So Out of Hand In America’s Schools That the Blue … https://www.reddit.com/r/Futurology/comments/1kztpy1/ai_cheating_is_so_out_of_hand_in_americas_schools/
Now let’s trace the claim: Berg links to a New York Magazine article stating, “rampant A.I. cheating has been reported all over the country”—can you follow this link (or search for the article by title) to read what evidence they actually provide, and summarize whether the term “rampant” is supported by specific data, examples, or just anecdotal teacher reports? (Share the article title, author, and your summary of what it says.) ✅
Great job gathering robust evidence! When you check the New York Magazine article Berg links for the claim “rampant A.I. cheating has been reported all over the country,” does the story cite specific studies, give data, or mostly use teacher and administrator anecdotes—and how does this affect your trust in the claim’s accuracy? (Be sure to share the title, author, and a short summary of what you found there.) ✅
on the uploaded document.
0 archived comments