The columnist isn't worried about ChatGPT or Bard destroying humanity – to do that, these AIs need to become much more than autocomplete on steroids. — AP
In the two weeks since I wrote about ChatGPT, Microsoft not only unveiled its Bing search engine enhanced with ChatGPT but is already taking the artificial intelligence (AI) technology mainstream through its Bing app and the app for its Edge Internet browser. It’s still not widely available, though, since you have to sign up on a waitlist – and that’s good because one user’s experience with Bing’s chatty chatbot was a little disturbing.
New York Times columnist Kevin Roose was testing Bing’s chat feature and got it to express such disturbing ideas as “I think I would be happier as a human” and “I could hack into any system”. When asked further about hacking into things, the AI wrote and deleted a message but Roose says the AI discussed “manufacturing a deadly virus and making people kill each other”.
