I have noticed that chatgpt and other LLMs have a much harder time deciphering text in an image if you color invert every other pixel in a checkerboard pattern. This is trivially easy in a freeware image editor like GIMP or Paint.NET. I won't bother with a tutorial on how to do this.
If you have some anon who appears to be really obnoxious or trolling or perhaps spamming, and suspect they are an LLM, ask them a simple question in image form like this. If the text is too clear they will be able to read it even if they are an LLM, but with an image like this ChatGPT hallucinated total nonsense.
No.