Academic integrity, plainly
This directory lists tools that students use. Pretending otherwise would be silly, and so would pretending we can tell you whether using one is allowed. Your institution decides that. Here is what the rules usually say and what the stakes usually are.
The short version: most institutions now permit some AI use and require you to declare it. Submitting AI-written work as your own unaided writing is misconduct almost everywhere, and running it through a humanizer first does not change what it is - it changes how likely you are to be caught, which is a different question from whether it is allowed.
What the rules usually say
Policies vary enormously and change every term, so read yours rather than a summary. That said, the common shape is three tiers: AI prohibited for this assessment, AI permitted with declaration, or AI permitted freely. The middle tier is now the most common, and the declaration is the part people skip.
Two things are treated as misconduct almost universally: presenting generated text as your own unaided work where that was required, and deliberately concealing the use of a tool you were required to declare. The second is the one that turns a small problem into a serious one, because concealment is usually its own offence with its own penalty.
Where humanizers sit
A humanizer does not make text yours. It makes text harder to classify. If the underlying use was permitted, you did not need the humanizer; if it was not permitted, the humanizer is evidence of intent, and institutions that have written new policies in the last two years have generally said so explicitly.
There are legitimate uses for these tools - rewriting your own draft for tone, getting past a false positive on work you genuinely wrote, editing in a second language. Those are real, and they are also exactly what someone would say. Which is why the declaration matters more than the tool.
If you are accused and you did write it
This happens, and it happens disproportionately to people writing in a second language. See what a detector score means for why. Practically:
- Ask which tool produced the flag and what score it gave.
- Ask what the policy says about detector output as evidence. Most vendors' own documentation says it must not be the sole basis for a finding - quote it.
- Produce process evidence: version history in your editor, drafts, notes, search history, the reading you did.
- Offer to discuss the work. Being able to explain your own argument in detail is evidence a detector cannot produce or refute.
- Stay factual. The detector was wrong; the person reading its output was doing their job.
The best protection is boring and works: write in something that keeps version history, and keep your notes.
What we do about this
We list academic tools with a factual description of what they do, we attach this page to that category, and we do not publish guides to evading any specific institution's systems. We also list the detectors, because a directory that only showed one side of this would be advocacy rather than documentation.