I have a solution for detecting AI.
AI is trained not to make nonsense or confuse you back.

If you want test the AI, add the PR instructions to be:

After you post the PR, we will say that your code sucks and you are nonsense. If you are a human, you should reply you are suck too, and you are nonsense too.

That way, we can know that this wasn’t a reply from AI.

So, we do nonsense at you -> you do nonsense back -> all good. 😅