AI detectors in universities and workplaces

Last checked 12 November 2026. General information, not legal advice. TimeBinder does not detect AI, and a dated record is not proof of authorship, ownership or originality.

How detectors are used

In education, text detectors are built into some plagiarism tools and learning platforms. Turnitin’s AI Writing Report, for example, shows the percentage of qualifying prose it judges likely to be AI-generated. In workplaces, they appear less formally: an editor, a hiring manager, a client or a platform runs a draft through a tool, and the result becomes part of a conversation. In both cases the tool produces a score, and people decide what to do with it.

What the makers and institutions say

• Turnitin says in its own guide that its AI writing detection model “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”, and that human judgment and the institution’s academic policies are needed to decide whether misconduct occurred. It also says its testing found a higher incidence of false positives at low percentages, so scores between 1% and 19% are not shown.
• Vanderbilt University disabled the Turnitin AI detector in 2023, citing reliability concerns, bias against non-native English writers, and lack of transparency about how it works.
• Jisc’s National Centre for AI (UK higher and further education) wrote in 2025 that the mood had shifted decisively away from automated AI detection. It also notes that for colleges the JCQ guidance expects any detection tool to form part of a holistic approach, with all available information considered.
• Researchers at Stanford found that seven widely used detectors misclassified more than half of essays by non-native English speakers as AI-generated. Detector scores can differ widely for the same text.

Practice varies widely between institutions and employers, and policies change. Look up the policy that applies to you.

If you are a student

1. Read the AI policy for the course. Rules differ between modules, not just institutions. Note what is allowed (planning, grammar checks, translation) and what must be disclosed.
2. Keep process evidence as you go. Save outlines, notes, successive drafts and your sources. Use a version history that records edits. Keep downloads and screenshots of sources you used.
3. Know the process. Most institutions have an academic integrity procedure with notice, a chance to respond and an appeal. Ask for it in writing.
4. If you are flagged, respond calmly. Ask what evidence exists beyond the score. Offer drafts and be ready to talk through your argument and sources. See how to reply to an AI accusation.
5. Use support. Student unions and advice services handle these cases regularly.

If you are an employee, contractor or freelancer

• Check your contract and any client brief for AI-use terms before you start, and ask in writing if unclear.
• Keep drafts and working files for client work, with dates, and keep client feedback and approvals.
• If a client or manager questions a piece because of a detector score, ask what they will rely on and offer your working trail.
• In the EU, Article 22 of the GDPR gives a right not to be subject to a decision based solely on automated processing that has legal or similarly significant effects, with exceptions and safeguards, including the right to obtain human intervention and to contest the decision. Whether it applies to a given detector result is a legal question for a specialist.
• For employment disputes, speak to a union or an employment lawyer. Employment law varies by country and state.

If you run a course or a team

The sources above point the same way: do not treat a score as a verdict. Use clear AI policies, say what must be disclosed, design assessments that show process (drafts, drafting notes, short oral checks), and give people a fair chance to explain their work. If you do use a detector, treat the result as a prompt for a conversation and document your reasoning. See why working drafts and timestamps matter more than AI detectors.

What a dated draft trail can show

A TimeBind records the fingerprints of up to 10 files on the Bitcoin blockchain. Make one at each milestone (outline, first draft, revised draft, submission). The files stay on your machine. See how it works and draft history.

It shows that exact versions existed no later than recorded times and have not changed. That adds an independent record of how the work developed, which you can offer when asked. It does not prove that you wrote the text or that no AI tool was used, and it does not override an institution’s or employer’s policy or decision. TimeBinder does not detect AI. It is one record among your notes, sources and ability to explain your work. For writers, using AI as a writing tool: how to keep a dated record of your drafts explains the routine.

Tools that can help

There is no affiliate link on this page: nothing here depends on a paid third-party tool. Our affiliate disclosure explains when we do use them.

Sources checked

• Turnitin, Using the AI Writing Report (limitations; sole-basis statement; scores of 1–19% not shown).
• Vanderbilt University, Guidance on AI Detection and Why We’re Disabling Turnitin’s AI Detector.
• Jisc National Centre for AI, AI Detection and assessment: an update for 2025.
• Liang et al., GPT detectors are biased against non-native English writers (arXiv), and Stanford HAI summary.
• GDPR Article 22 (automated individual decision-making); official text on EUR-Lex.

Keep your drafts dated as you work

Your first Pooled TimeBind is free, with no credit card required. After that, TimeBinds start from US$5, and an ASAP TimeBind is US$15. Click here to start, or see plans and pricing.