Can AI detect each other?

That fast! Stanford researchers just launched zeroGPT (https://www.zerogpt.com/), which they claim is "the most advanced and reliable ChatGPT detector tool". It claims to be able to identify content generated by gpt. So here is the problem: can AI detect each other? There are three views: 1. AI cannot detect AI According to the Turing test, AI is indistinguishable from humans. 2. AI can detect AI Because the content generated by AI is very repetitive, there are patterns that can be identified. 3. Prompt related If the written prompt is very general, it may be recognised; if it is very specific and detailed, it may not be recognised. What do you think? Please comment with your insight!