🔬 Science and technology · Intermediate · Policy resolution

Resolved: High schools should allow students to use generative artificial intelligence tools for schoolwork.

Most teens now use chatbots for schoolwork, with or without permission. Should schools allow it, and teach students to use it well, or restrict it?

Background

Pew Research Center found the share of U.S. teens (13–17) who had used ChatGPT for schoolwork doubled from 13% in 2023 to 26% in 2024; teens were far more likely to call it acceptable for researching new topics (54%) than for writing essays (18%). A fall 2025 Pew survey found about 54% of teens used chatbots for schoolwork, and 59% said cheating with these tools had become a regular part of student life. [1][2]

Early experiments point both ways. In a randomized trial in a Harvard introductory physics course (Kestin and colleagues, 2025), students learned more from a carefully designed tutoring tool than from an in-class active-learning lesson on the same topic. In a randomized trial with nearly 1,000 high school math students in Turkey (Bastani and colleagues, PNAS 2025), students with an unrestricted chatbot did better during practice but 17% worse on a later exam without it, while students using a version built to give hints rather than answers did about as well as students without the tool. [3][4]

The difference between the two studies points to the real question: not only whether students use these tools, but how. Schools’ policies range from bans to rules that allow specific uses (brainstorming, feedback, practice) while forbidding others (submitting generated work as one’s own). [3][4]

Where the facts themselves are disputed

Do these tools help or hurt learning?

  • Some controlled studies find well-designed tutoring tools raise learning, sometimes substantially.
  • Others find unrestricted use acts as a crutch: students do better while using the tool and worse without it.

Key terms

Generative artificial intelligence
Software that produces text, images or other content in response to prompts, such as chatbots.
Guardrails
Design limits on a tool, such as giving hints instead of full answers.
Academic integrity
Honesty in schoolwork: doing your own work and crediting sources.

Both sides at their strongest

Each argument is written the way its best advocates would make it, followed by the strongest answer from the other side. Use them to build your case and to prepare for the other one.

Affirmative

Schools should allow these tools and teach students to use them well, because students will use them anyway and can learn more with them.

  1. A1 Well-designed tools can raise learning.

    Why A tutor that explains step by step, answers questions instantly and adapts to each student offers help that a teacher with thirty students can’t always give.

    Evidence Kestin and colleagues (2025, Scientific Reports): students learned more with a structured tutoring tool than in an active-learning class session.

    Best answer The negative answers that the same evidence shows design matters: unrestricted tools can harm learning, as Bastani and colleagues found.

  2. A2 Students already use them; schools should teach responsible use.

    Why Bans push use out of sight. Allowing use with clear rules lets teachers teach when it helps, when it doesn’t, and how to check its errors.

    Evidence Pew surveys showing most teens use chatbots for schoolwork; school district policies that permit specific uses.

    Best answer The negative replies that widespread misuse is a reason to restrict, not to permit, and that rules for responsible use are hard to enforce.

  3. A3 It helps students who lack other help.

    Why Students whose families can’t pay tutors or help with homework gain access to explanations at any hour, narrowing gaps.

    Evidence Studies of tutoring’s effects on achievement; reports on access to private tutoring by family income.

    Best answer The negative argues that students with less support may also be most likely to lean on the tool as a crutch.

  4. A4 Using these tools is a skill students will need.

    Why Many jobs and colleges now expect people to use these tools; learning to use them critically is part of preparing for that.

    Evidence Employer surveys; college policies allowing guided use.

    Best answer The negative replies that students need strong basic skills first, and the tools change so fast that specific skills date quickly.

Negative

Schools should restrict these tools for schoolwork, because they let students skip the thinking that learning requires.

  1. N1 Unrestricted use can harm learning.

    Why Learning comes from struggling through problems. If a tool supplies answers, students practice copying, not thinking.

    Evidence Bastani and colleagues (PNAS 2025): students with an unrestricted chatbot scored 17% lower on a later exam without it.

    Best answer The affirmative answers that the same study found a hint-based version didn’t harm learning, so the answer is rules about design and use, not restriction.

  2. N2 It makes cheating easy and hard to detect.

    Why Generated essays and answers are hard to tell apart from student work, which undermines grades and trust.

    Evidence Pew (2025): 59% of teens say cheating with these tools is a regular part of student life; research on the unreliability of detection software.

    Best answer The affirmative replies that cheating happens under bans too; changing assignments (in-class writing, oral checks) works better than prohibition.

  3. N3 The tools make mistakes students can’t catch.

    Why Chatbots can produce confident but wrong answers. Students who are still learning a subject are the least able to notice.

    Evidence The Bastani study’s finding that the unrestricted tool’s answers were correct only about half the time; reports of fabricated citations.

    Best answer The affirmative argues that teaching students to check answers is itself valuable, and newer tools with guardrails make fewer errors.

  4. N4 Privacy and fairness concerns remain.

    Why Students’ questions and work are sent to private companies, and not every family can afford the best versions.

    Evidence Student-privacy laws such as FERPA and COPPA; school district reviews of vendor contracts.

    Best answer The affirmative replies that schools can choose approved tools with privacy protections and provide them free to all students.

Strawman or steelman?

A strawman is a weak copy of the other side’s view, easy to knock down and unfair to the people who hold it. A steelman is that view at its strongest. Answer the steelman, and you’ve answered the real argument. (More on steelmanning.)

Affirmative

Strawman “They want students to let a chatbot do all their homework.”

Why it’s unfair Supporters favor allowed uses with rules, not handing over the work.

Steelman Students already use these tools; schools that permit specific uses can teach responsible habits and use well-designed tutors that research shows can raise learning.

Negative

Strawman “They’re just afraid of new technology.”

Why it’s unfair Opponents point to controlled evidence that unrestricted use harms learning, plus cheating and accuracy problems.

Steelman Learning requires doing the thinking; unrestricted tools let students skip it, as experiments show, and they make cheating easy and errors hard to catch.

Weak arguments to avoid

Affirmative

  • “Chatbots are always right.” They make frequent errors; claiming otherwise is easily refuted.

Negative

  • “Technology always makes students dumber.” Overgeneralized; calculators, libraries and the internet have mixed and often positive effects.

Questions for cross-examination

Ask the Affirmative

  1. What happened in the Bastani study when students used an unrestricted chatbot?
  2. How would a teacher know whether a student wrote an essay?
  3. Which uses would you forbid?

Ask the Negative

  1. What did the Harvard physics study find?
  2. If most students already use these tools, does a ban work?
  3. Would you allow a hint-only tutoring tool?

Reading on both sides

Read at least one source from each side before you debate. Look up each source’s author and publisher too: who wrote it is part of the evidence.

Background and overviews

Mostly Affirmative

Mostly Negative

Practice this topic

Sources for the background

  1. More Teens Are Using ChatGPT for Homework (on Pew Research Center data), Britannica ProCon (2025-01-15). Accessed 2026-10-25.
  2. Survey finds 54% of U.S. teens use AI for schoolwork, GovTech (on Pew Research Center data) (2026-02-26). Accessed 2026-10-25.
  3. AI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting, Scientific Reports 15:17458 (2025). Accessed 2026-10-25.
  4. Unrestricted generative AI harms high school math learning by acting as a crutch, PsyPost. Accessed 2026-10-25.

Teddy Rook Book doesn’t take a side on this resolution. Updated 2026-10-25.

@teddyrookbook GitHub