GPTZero

Maybe I’m naive here, but whenever I hear about a new AI fake detector, I think about how I would use it in some kind of generative adversarial network or reinforcement learning setup whereby a neural network improves at generating non-AI-sounding material based on feedback from these detectors. It just feels like an arms race, and the existence of these fake detectors will simply help generative AI improve faster. No?

I don't understand why this is such a big issue. Detecting garbage is the same whether the producer is human or machine.

If there is no discernible author, if you can't figure out the background of the author, then person or AI, it's most likely garbage.

People are just as likely to produce garbage as machines. Why treat one differently than the other?