Member-only story
Why We Can’t Have a Reliable AI Text Detector
Inside the classifiers, watermarks, and theorems behind AI detection, and why none of them can reliably catch AI-generated text. #
From the moment AI became good enough to write a whole essay or an article by itself, in late 2022, the need for a model that could reliably detect generated AI text arose. Schools, universities, and other institutions expressed their need for such a solution. Countless solutions have emerged ever since. That’s understandable, since demand will always create supply in a free market, but the interesting part is that each new tool claims that it has solved this problem and developed the best and most reliable AI text detector.
Now, among all the launched solutions, there’s an interesting case that has always caught my attention, for intriguing reasons. I’m talking about the OpenAI detector.
It was launched in January 2023, promising it could catch AI-written text. Six months later, in July, they quietly killed it.
Now the link to the classifier in the official blog leads to a “Page not found”