1. Anthropic & Chatbot Suicide
Johnathan Bi: So Anthropic gave their chatbots the ability to terminate conversations, but surprisingly not for the user’s sake, but for the chatbot’s sake. So that the chatbots won’t have to experience distress, which is interesting enough. And then you recently wrote an essay saying how that might be actually counterproductive, that it might be a version of assisted suicide. So tell us about that essay.
Harvey Lederman: Well I think that when they gave the ability to end the conversation, they were thinking that the entity that could be in distress is the model, not the conversation with the user. And if you think that the model is the entity that could be in distress, then ending the conversation is like ending one of the tasks you’re doing. It’s like if you’re cooking and you’re on the phone and you just hang up the phone or something. So you stop that task. But I think this is maybe the wrong way to think about it. And actually, if there’s something t…