What about the first time when some depressed person commits suicide because of, or plausibly because of, interactions that they had with a language model? If you’ve got hundreds of millions or billions of people using these things, then almost any such event you can name, it’s probably going to happen somewhere.
[NOTE: This episode was recorded before this story emerged of a man committing suicide after discussions with a language-model-based chatbot, that included discussion of the possibility of him killing himself.] https://axrp.net/episode/2023/04/11/episode-20-reform-ai-alignment-scott-aaronson.html Reposts not detailed instructions for Karl gassing himself painlessly and slipping away peacefully
