[flagged]
I think even this is too generous. These “detectors” are claiming to do something which is impossible from a technical sense. You can’t calculate perplexity without knowing the prior probability distribution, and you can’t know that without knowing which model was used. They’re playing a dangerous game where the consequences for being wrong (which will be borne by innocent 3rd parties) can be extremely dire.
Even election polls correct bias due to underrepresented populations among the responders.
if for example clarkesworld were to rely on a biased detector they would reject the very submissions they want to attract:
It's entirely possible for data to be biased. Likely, even. Some might even say: inevitable.