Start here
The free tool below runs entirely in this browser — no account, nothing stored or transmitted. Redact anything you would not want kept before you type it.
Open the companion tool →Why the record beats the argument
Detection tools estimate; they do not know. A score is a probability produced by a model with its own error rate, and schools that rely on one often cannot explain how it was produced. Asking for the basis is a reasonable, non-accusatory question.
What changes outcomes is documentation of the process: draft files, version history, notes, a library receipt at a specific hour, the teacher's earlier feedback, or a writing sample from before the assignment. A record outlasts any argument about tone.
The protocol
- Write the timeline first: when the assignment started, when work happened, which files or apps were used, and when it was submitted.
- Collect the artifacts that show process rather than product: version history, drafts, notes, browser history for research, a photo of handwritten work.
- Write one calm sentence about the tool: detection tools estimate and produce errors in both directions; you are asking how this decision was made, not who is lying.
- Make one specific request in writing: the basis for the finding, the tool or method used, and how it can be reviewed under the school's policy.
- Ask about the process going forward — a documented alternative like an in-class writing sample, which resolves the question for everyone.
What the tool does
Build a written response to an AI-detection accusation: the timeline, the evidence that exists, and one request for the basis of the decision. It does not accuse anyone and it does not score the work.
Limit first: This is a response-preparation aid, not legal advice and not a judgment about whether work was written with AI. School policies and appeal routes differ by district and country.
Open the companion tool →One human next step
Send the written response to the teacher and, if unresolved, to the school's administration in the order the policy requires, keeping copies of everything. If a decision stands on an unexplained automated score, a school-board or district-level process is the next step.
Research log and safety checks
AI-assist path: not used for a verdict. Both halves of this topic — the accusation and the defense — are probabilistic judgments, so §1.2 UIC-10 applies: the artifact asks for the basis of a decision and assembles the student's own record instead of scoring anything. §6.10: not a frontier-pace topic.
§0.5 protection result: category is family-teen-safety; crisis-adjacent: no; fear/urgency/scarcity toolkit used: no. The protective function is free and the paid feature is convenience only.