Comment by Noelia-

This kind of thing makes me a bit cautious. When you use ChatGPT a lot, it’s easy to read too much into its replies. The model is just predicting text, but people bring their emotions into the conversation. For vulnerable users, that can spiral quickly.

The real risk isn’t the AI itself, but how people interact with it. We definitely need stronger safeguards to keep more people from getting pulled in too deep.