AI Detection Tools
5 tools listed in this category, compiled by hand from each provider’s official site and documentation.
AI detection tools estimate the probability that a piece of text, or in some cases an image, was produced by a generative model. They do this by measuring statistical properties of the writing — how predictable each word is given the ones before it, and how much the sentence length and structure vary. Text produced by a language model tends to be more uniform on these measures than text written by a person in a hurry.
That method has a hard limit, and it is the most important thing to understand before using any tool in this category. A detector produces a probability, not proof. Human writing that is plain, formulaic or written by someone in a second language can score as machine-generated, and machine text that has been edited by a person can score as human. Published error rates are not zero, and they are measured on test sets that may not resemble your material. No score in this category should be treated as evidence on its own.
The practical consequence is about process rather than product. Using a detection score to accuse a student or reject a freelancer, with nothing else to go on, is unfair and frequently wrong. Used properly, a high score is a prompt to look further — at draft history, at version records, at a conversation about the work — not a verdict. If you are setting policy, write down what happens after a flag before you buy anything.
Reasonable starting points from this list: GPTZero and Copyleaks both offer free tiers and are the simplest way to see how variable results are on text you already know the origin of. Originality.AI, paid, is aimed at publishers and agencies checking commissioned work at volume. Hive Moderation covers image and media checks in addition to text. Our separate guide explains how to read these results in more depth.
