> But I always wonder if there would be moments where someone writes a genuine thing and it’s mistaken for AI.
That certainly must happen, because the classifiers are imprecise.
So far, though, each time a case of possible misclassification has come up on HN and I've worked with the user to figure out what happened, it turns out that they used tools that they didn't think of as genai, but which left LLM imprints on their text. More about this at https://news.ycombinator.com/item?id=48405497.
This evidence is a bit weak, though, because the sample size is small.
That certainly must happen, because the classifiers are imprecise.
So far, though, each time a case of possible misclassification has come up on HN and I've worked with the user to figure out what happened, it turns out that they used tools that they didn't think of as genai, but which left LLM imprints on their text. More about this at https://news.ycombinator.com/item?id=48405497.
This evidence is a bit weak, though, because the sample size is small.