New AI classifier for indicating AI-written text

We’re launching a classifier trained to distinguish between AI-written and human-written text.

Read more in our newest blog post and check the free tool out for yourself!


I tested but no good I’d say. Because in almost all cases it told me unable to determine. I had provided 1000 or so characters as was indicated in the blog for best results


Hi Logan, can I check if the tool has been removed? Because it’s showing Page Not Found for more than 3 days already.

1 Like

Hi Sam14, this does indeed seem to have been removed as a service.

Hi @Foxalabs I was hoping for official confirmation because we used this for our work. Do you know if we will get official confirmation?

You can reach out via email to for clarification, but I think it fairly likely that’s it’s now depreciated.

@Foxalabs Thanks for the info. I appreciate that!

I’m sure you have your reasons, but this seems crazy to me given the actual text from OpenAI announcing the classifier,

Our classifier is not fully reliable. In our evaluations on a “challenge set” of English texts, our classifier correctly identifies 26% of AI-written text (true positives) as “likely AI-written,” while incorrectly labeling human-written text as AI-written 9% of the time (false positives). Our classifier’s reliability typically improves as the length of the input text increases. Compared to our previously released classifier, this new classifier is significantly more reliable on text from more recent AI systems.

But to address your concern,

As of July 20, 2023, the AI classifier is no longer available due to its low rate of accuracy. We are working to incorporate feedback and are currently researching more effective provenance techniques for text, and have made a commitment to develop and deploy mechanisms that enable users to understand if audio or visual content is AI-generated.


This is inherently a mission: impossible scenario. Text generated by AI is sourced from human written text. There are countless numbers of combinatorial sequences of sentences any human could generate in equally countless number of possible scenarios. Take, for instance, the possibility that a humans output could in itself be influenced by an AI’s outputs over time - there exists no methods to diferentiate, conclusively, between the two.

AI generated imagery, audio, etc - that is a different thing all together and your approach I am sure is much more feasible in those applications at least at this point in time along in their evolution.