Comment by Eextra953
17 hours ago
I've looked at this same problem in an academic environment and have come to conclude that there is no way to reliably detect AI writing using only text. The reason for this is that no detector can take two identical inputs and classify one as synthetic (LLM) and one as organic (Human) and this situation can easily happen at the sentence or even paragraph level. Posed as a question, if a human writes a paragraph that happens to be exactly the same as a paragraph written by an LLM how do you classify that paragraph? There are a finite number of words and a finite way of combining them within an academic setting/field which means that we can't build a perfect classifier with just text. In practice, there are obvious tells and LLM-isms but these also change with time and each model has different tells so that even if a paper is full of LLM like writing we have no way to disambiguate between organic text that appears synthetic or synthetic text that appears organic.
A more interesting question, to me, is looking at a corpus of essays and analyzing how writing has changed with the introduction of LLMs. We can look at changes in vocabulary, linguistic features, style embeddings, regular embeddings, typos, errors, and references over time. When looked at in this way it is clear that academic writing has changed at the population level but what has led the change is harder to track down.
When some closed models are "retired", interestingly we might not even be able to extract all their tells. We'd be limited to people writing about them on forums (where they don't often specify version, and there's a lot of memes and confirmation bias) and I don't know, doing some statistical Bayesian guessing that given the likelihood it's FooGPT 10.5, maybe we should update our beliefs about its characteristics etc.
> There are a finite number of words and a finite way of combining them within an academic setting/field which means that we can't build a perfect classifier with just text.
Maybe for very short phrases, but otherwise I disagree. Phrasing very quickly runs into a combinatorial explosion. In the words of Noam Chomsky, "Virtually every sentence that a person utters or understands is a brand-new combination of words, appearing for the first time in the history of the universe."
In my opinion, the difficulty in LLM/human text discrimination isn't that a person might coincidentally write exactly the same text as an LLM would, but rather that 1) LLMs aren't hard locked to a single phrasing (so this is a tougher problem than matching to a single static document, e.g. plagiarism detection) and 2) text has relatively low information density, so you need quite a bit of it to gather enough data to run a statistical test with a reasonably narrow confidence interval.