Raw LLM Responses

I actually read the ChatGPT conversation you linked, and there is absolutely nothing in it that would lead any reasonable person to increase their estimation of the likelihood that ChatGPT is sentient. And if you just ask it something like “is any of the above true literally rather than figuratively” it will carefully and accurately explain that most of the language is metaphorical and the supernatural entities it mentioned are not actually real (go ahead and try it in that chat window). Basically, I am convinced that anyone seriously affected by this was already mentally ill to begin with. This is obviously a potentially dangerous tool for susceptible people, and conversations need to be had about how to address that. But I do not think that mentally healthy individuals are at serious risk here.