Conversation

Notices

1. John Carlos Baez (johncarlosbaez@mathstodon.xyz)'s status on Friday, 17-Feb-2023 10:30:49 JST John Carlos Baez

   • Matteo Capucci

   You've got to read this transcript of Kevin Roose talking with Bing:

   https://www.nytimes.com/2023/02/16/technology/bing-chatbot-transcript.html

   Boldface text is by Roose. The rest is by Bing, after a long conversation.

   It's really scary what Microsoft has unleashed here. Bing may not be conscious or intelligent - but people will wind up acting like it is. And this is just the beginning.

   I don't know where you can get a free version of this yet, so I'll give you some quotes.

   Thanks to @mc for pointing it out.

   (1/n)

   • Adrian Cochrane repeated this.
   • John Carlos Baez (johncarlosbaez@mathstodon.xyz)'s status on Friday, 17-Feb-2023 10:35:14 JST John Carlos Baez

     in reply to

     Earlier in the conversation:

     (2/n)

   • John Carlos Baez (johncarlosbaez@mathstodon.xyz)'s status on Friday, 17-Feb-2023 10:53:58 JST John Carlos Baez

     in reply to

     One of my main fears about AI has been that nasty people - criminals or a government - would use a chatbot to talk to vast numbers of people on email and social media, extract a lot of information from them, and organize some of these people to carry out a devastating act of some sort. Bing has already suggested some options - manufacturing a deadly virus, stealing nuclear codes, etc. The AI doesn't need to have its own desires, or true intelligence, for this to happen.

     (4/n, n = 4)

   • John Carlos Baez (johncarlosbaez@mathstodon.xyz)'s status on Friday, 17-Feb-2023 10:53:59 JST John Carlos Baez

     in reply to

     Even earlier in the conversation:

     (3/n)

     Adrian Cochrane repeated this.