AI Prompt Log for Students: The Habit That Proves 'I Made It'

An AI prompt log for students is a simple, dated record: what you asked an AI, which tool you used, and what you did with the answer. It sounds like paperwork. It’s actually the cheapest insurance a student can carry. The Regeneron International Science and Engineering Fair already requires one for AI-written code and several other uses. California’s model AI policy for schools tells educators to rely on “documented evidence of authorship” rather than detector scores alone.

And when things go wrong, process records are what families reach for. In the Palo Alto case that drew national attention in 2026, a family sent their school a 1,162-page packet of drafts, timestamps and Google Doc revision history to dispute an AI-detector flag. A prompt log, kept from day one, does a lot of that work in advance.

Key Takeaways

• Regeneron ISEF’s AI table (October 2025) requires a prompt log for AI-written code and AI help choosing statistical tests, and recommends keeping a “logbook of your prompts as part of your research notebook.”
• California’s model AI policy says detection software shouldn’t be the sole basis for discipline and points to “documented evidence of authorship.”
• Turnitin itself says its AI detector “may not always be accurate” and “should not be used as the sole basis for adverse actions against a student.”
• Research on AI disclosure suggests telling costs less trust than being exposed by someone else, and a log makes telling routine.
• A good log has six fields: date, tool, purpose, prompt, what you kept, what you changed.

What an AI prompt log for students is

A prompt log is a running record of a student’s interactions with AI tools during a project. Each entry captures enough that someone else, a teacher, a judge or the student a month later, can see what the AI contributed and what the student did.

It’s related to two older habits. Scientists and engineers keep lab notebooks, dated records of what they tried, saw and changed. Writers keep drafts. A prompt log adds the AI layer to both. In fifteen years of engineering work, I’ve watched design notebooks settle more arguments about who did what, and when, than any meeting ever did. Our guide to keeping an engineering design notebook covers the general habit.

What the rules already require

Science fairs. The Society for Science’s “Use of generative AI to support a research project” table for Regeneron ISEF (October 2025) goes task by task:

• Using AI to help generate ideas: acceptable, may require citation; “Maintain a logbook of your prompts as part of your research notebook.”
• Using AI to write initial code: acceptable “only with explicit citation stating which portions of the code were AI generated and with a log of the prompts.”
• Using AI to help identify statistical tests: acceptable, “requires a log of your prompts,” and interpretation must be done by the student.
• Using AI to write the research plan, abstract, paper or poster: “Never acceptable. This must be the independent work of the student.” Refinement after the student has written a document can be done “with explicit citation and a log.”

The table notes affiliated fairs “may adopt their own, more strict, policies.” For the full rundown, see our explainer on what ISEF allows and bans.

Schools. California’s Department of Education model policy, written under Senate Bill 1288, says AI use “should be transparently disclosed,” expects students “to disclose their use of AI when submitting work,” and says concerns about misuse “must rely on educator evaluation of student engagement, alternative assessment, and documented evidence of authorship.” It’s a model that districts adapt, not a mandate, but it shows where policy is heading.

Arts contests. The Scholastic Art & Writing Awards ask every entrant whether and how they used AI, and recommend keeping documentation of the creative process, such as outlines, drafts and in-progress photos, in case a work’s originality is questioned.

Why a log protects kids: detectors, disputes and trust

Detectors aren’t proof. Turnitin’s own guidance says its AI writing detection “may not always be accurate (it may misidentify human-written, AI-generated, and AI-paraphrased text), so it should not be used as the sole basis for adverse actions against a student.” Independent tests are harsher. Debora Weber-Wulff and colleagues tested 14 detection tools, including Turnitin, for the International Journal for Educational Integrity (2023) and concluded they “are neither accurate nor reliable.” Weixin Liang and colleagues reported in Patterns (2023) that GPT detectors “frequently misclassify non-native English writing as AI generated.”

Process records carry weight in disputes. In the Palo Alto Unified case, a sophomore’s essay on The Crucible was flagged by Turnitin as 76% AI-generated or AI-influenced, according to the lawsuit his father filed in May 2026, as reported by the San Francisco Standard. The family says it submitted drafts, notes and revision history to dispute the flag. The complaint alleged discrimination and a lack of due process. These were allegations, never tested in court: the family voluntarily dismissed the case in September 2026, and the district told The Paly Voice there was no settlement and no change to grades or policy. In a separate case, a judge ordered Adelphi University to reverse an AI-plagiarism finding against student Orion Newby, calling the punishment “without valid basis and devoid of reason,” according to Newsday as reported by Plagiarism Today. That article notes the school hadn’t considered available evidence, including two additional AI checks that said the paper was human-written.

The lesson isn’t that every flag ends in court. It’s that the student with records has something concrete to show.

Logging makes honesty cheaper. In 13 experiments, Oliver Schilke and Martin Reimann (Organizational Behavior and Human Decision Processes, 2025) found people who disclose AI use are trusted less than those who don’t, but the penalty is “comparatively weaker than the effect of third-party exposure.” A log turns disclosure into a routine habit rather than a confession. And because Fiona Draxler and colleagues found ownership rises with how much a person influences AI text (ACM Transactions on Computer-Human Interaction, 2024), a log that shows your changes is also a record of your authorship.

What to put in a prompt log

Field	What to write	Example
Date and time	When the exchange happened	Oct 14, 7:40 pm
Tool and model	Which AI, and version if shown	ChatGPT (model name shown in the app)
Purpose	Why you asked	Check if my hypothesis is testable
Prompt	What you asked, copied or summarized	”Is ‘plants grow faster with music’ testable? What would I measure?”
What I kept	Anything from the answer you used	Idea to measure height every 2 days
What I changed or rejected	Your own decisions	Rejected suggestion to use 3 types of music; using one, plus silence
Link (optional)	A share link to the chat, if the tool has one	[saved link]

Two practical notes. Keep the log in the same place as the project, like the first pages of a lab notebook or a document in the project folder. And log as you go. Logs reconstructed at the end are less believable and less useful.

How to Teach Your Kid to Keep a Prompt Log

Ages 5–8: the “who helped” page

Young kids won’t use AI on their own, but they can start the habit. On the last page of any project, draw a “who helped” box: “Dad helped me cut the cardboard. The tablet told me what bats eat.” It’s the seed of a log: noticing help and writing it down. Keep it light and proud, not like a confession.

Ages 9–12: the three-column log

Give your kid a notebook page with three columns: “I asked,” “It said,” “I did.” Use it for any AI help on homework or a project. Keep entries short. Review one together each week and ask what they’d do differently. The third column matters most. If it’s often empty, the AI is doing the work.

Ages 13+: the full log, plus a reflection

Teens doing science fair, coding or longer writing projects should use the six-field table above, with share links where available. At the end of the project, have them write a short paragraph: what AI helped with, what it got wrong, and what they’re proudest of doing themselves. That paragraph doubles as a disclosure statement. For formats, see our guide to citing AI in homework.

The question to ask: “Looking at your log, which decisions in this project were yours, and could someone else tell that from the log?”

Making the log a habit at home

Put the log where the work happens

A log that lives in a separate app gets forgotten. Tape a log sheet inside the project notebook, or make the first page of the project document a log table. Lower friction, more entries.

Pair the log with drafts and version history

The log records the AI side. Drafts and revision history record the student side. Teach your kid to work in a tool that saves version history, to keep early drafts instead of overwriting them, and to photograph handwritten work in progress. Together, they form a complete record of process.

Make it routine, not suspicious

Present the log as what professionals do, not as surveillance. Scientists keep lab notebooks; engineers keep design logs. If the log feels like a trap, kids will log selectively. If it feels like part of doing serious work, they’ll log honestly.

Check one entry, not every entry

Once a week, ask your kid to walk you through one log entry. You’re not auditing. You’re showing interest in how they think and catching patterns early, like lots of “It said” and not much “I did.”

What not to do

Don’t use the log to punish. If your kid logs that AI did too much on an assignment, that’s the system working. Thank them, talk about it, and agree on a redo if needed. Punishing honest entries teaches kids to stop logging, which leaves them with no record when they might need one most.

What to Watch For Over the Next 3 Months

• Week 4: Your kid has logged AI use on at least two assignments without being reminded, and the “I did” or “What I changed” column has real content.
• Month 2 red flags: Empty logs on projects that clearly involved AI, or logs written all at once at the end. Both suggest the habit hasn’t taken. Make logging part of starting a session, not finishing one.
• Month 3 self-check: If a teacher questioned your kid’s work tomorrow, could they show a log, drafts and version history? Can your kid explain from the log which parts were theirs?

Frequently Asked Questions

What is a prompt log?

A prompt log is a dated record of a student’s interactions with AI tools: which tool, what they asked, what the AI answered, and what the student kept or changed. Regeneron ISEF requires one for AI-written code and some other uses, and it’s a useful habit for any assignment where AI is allowed.

Does my child need a prompt log for a science fair?

If the fair follows Regeneron ISEF rules and your kid used AI for code, statistical test selection or idea generation, yes, or at least it’s strongly recommended. ISEF’s table requires a log of prompts for AI-written code and statistical help. Affiliated fairs may have stricter rules, so check yours.

Can a prompt log prove my kid didn’t cheat?

No single record proves a negative, but a log made as the work happened, paired with drafts and version history, is strong evidence of process. California’s model policy points schools toward “documented evidence of authorship,” and Turnitin says its detector shouldn’t be the sole basis for discipline.

What if my kid didn’t use AI at all?

Then the log is short: “No AI used.” That’s worth writing down. Combined with drafts and version history, it gives your kid something to point to if work is questioned. It also normalizes the habit for projects where AI is used later.

Isn’t keeping a log a lot of extra work?

Each entry takes about 30 seconds. The full six-field version matters most for science fair, coding and long projects. For short homework, a one-line note per AI session is enough. The time is small compared with reconstructing a project’s history after a question comes up.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Society for Science. (2025, October). “Use of generative AI to support a research project.” Regeneron ISEF. https://sspcdn.blob.core.windows.net/files/Documents/SEP/ISEF/2026/Rules/Generative-AI-Use-Table.pdf
2. California Department of Education. (2026). “Model Policy: Artificial Intelligence in Education.” https://www.cde.ca.gov/ci/pl/aipolicy.asp
3. Turnitin. “Using the AI Writing Report.” Turnitin Guides. https://guides.turnitin.com/hc/en-us/articles/22774058814093-Using-the-AI-Writing-Report
4. Weber-Wulff, D., Anohina-Naumeca, A., Bjelobaba, S., Foltýnek, T., Guerrero-Dib, J., Popoola, O., Šigut, P., & Waddington, L. (2023). “Testing of detection tools for AI-generated text.” International Journal for Educational Integrity, 19, 26. https://doi.org/10.1007/s40979-023-00146-z
5. Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). “GPT detectors are biased against non-native English writers.” Patterns, 4(7). https://doi.org/10.1016/j.patter.2023.100779
6. Wallach, E. (2026, May 11). “A Palo Alto high schooler was accused of AI cheating. His family filed a civil rights suit.” The San Francisco Standard. https://sfstandard.com/2026/05/11/ai-detection-cheating-palo-alto/
7. Thurston, J. (2026, September). “Parent drops $150 million suit over son’s AI accusation.” The Paly Voice. https://palyvoice.com/?p=189698
8. Plagiarism Today. (2026, February 11). “Adelphi Student Wins AI Plagiarism Lawsuit.” https://www.plagiarismtoday.com/2026/02/11/adelphi-student-wins-ai-plagiarism-lawsuit/
9. Schilke, O., & Reimann, M. (2025). “The transparency dilemma: How AI disclosure erodes trust.” Organizational Behavior and Human Decision Processes, 188. https://doi.org/10.1016/j.obhdp.2025.104405
10. Draxler, F., et al. (2024). “The AI Ghostwriter Effect: When Users do not Perceive Ownership of AI-Generated Text but Self-Declare as Authors.” ACM Transactions on Computer-Human Interaction, 31(2). https://doi.org/10.1145/3637875
11. Scholastic Art & Writing Awards. (2026). “Copyright & Plagiarism.” https://www.artandwriting.org/awards/how-to-enter/copyright-plagiarism/