AI Content Detectors

Tools that estimate whether text or images were machine-generated.

3 tools

These tools estimate whether a piece of text or an image was machine-generated. It is worth being direct about this category, because the marketing is confident and the underlying reliability is not: no detector can determine this with the certainty its percentage score implies, and treating that number as evidence has already caused real harm to real people.

What to look at when choosing

The vendor with the most credibility withdrew its own detector

OpenAI released an AI text classifier and shut it down in July 2023, citing its low rate of accuracy. The organisation with the deepest knowledge of how the text was generated could not make detection work well enough to keep offering it. That is the most informative single fact about this category, and it is worth weighing against any vendor claiming high accuracy today.

False positives are not evenly distributed

Detectors key on statistical regularity — even sentence length, common word choices, predictable structure. That describes machine output, and it also describes careful writing by people composing in a second language, or anyone taught to write plainly. The people most likely to be wrongly flagged are the ones least equipped to contest it, which turns a technical limitation into a fairness problem.

A score is a signal, never a verdict

If you use one of these at all, use it to decide where to look more closely, not to decide an outcome. Version history, drafts, the ability to discuss the work — these are evidence. A percentage from a model that cannot explain its reasoning is not, and no institution should be attaching consequences to it on its own.

When you probably do not need one

If the underlying question is whether someone understands their own work, asking them about it is faster, fairer and far more conclusive than any detector. Most of the situations that send people to this category are better answered by changing the task than by buying a tool.