> So did the user. If he didn't want to talk to a chatbot he could have stopped at any time. This sounds similar to when people tell depressed people, just stop being sad.
IMO if a company is going to claim and release some pretty disruptive and unexplored capabilities through their product, they should at least have to make it safe. You put a safety railing because people could trip or slip. I don't think a mistake that small should be end in death.