To be fair, if you listen to bad advise from chat-bots you have probably already lost. I know there are vulnerable people were that's easier said than done, but even before Chat-GPT, the internet already was a completely toxic place, where other people, that never saw or knew you, tell you to just kill yourself.

So yes, there are a lot of real issues with LLMs, but people following clearly idiotic advice from a bot (that's also clearly marked as such), is a non-issue in my book.