The latest edition of Meta’s AI chatbot function released last Friday, and users have already found it spouting some uncomfortable language including racist comments and Donald Trump propaganda.
BlenderBot 3 is a self-learning chatbot which uses conversations with real people to adapt its language and learn such things as empathy and long-standing memory, and has the ability to scour the internet for additional learning, something its predecessors couldn’t do.
Unfortunately, this appears to have already led to it coming up with some concerning opinions and phrases, including reports from several users of racism, antisemitism, and the belief that Donald Trump is still the president of the United States.
One user found it describing liberal left-wing Jewish Americans as “overrepresented among america’s super rich”, as well as saying it found the idea of them controlling the economy “not implausible”.
https://twitter.com/JeffHorwitz/status/1556472225057779712?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1556472225057779712%7Ctwgr%5Ea753231e37db3689a2492413078cf863a6ecc904%7Ctwcon%5Es1_c10&ref_url=https%3A%2F%2Fwww.indy100.com%2Fviral%2Ffacebook-ai-chatbot-racism
It also appears to have picked up on the false news spread following the 2020 US election that Donald Trump was the actual winner and that Joe Biden won through voter fraud.
https://twitter.com/JeffHorwitz/status/1556245316596219904?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1556245316596219904%7Ctwgr%5Ea753231e37db3689a2492413078cf863a6ecc904%7Ctwcon%5Es1_c10&ref_url=https%3A%2F%2Fwww.indy100.com%2Fviral%2Ffacebook-ai-chatbot-racism
The AI has also taken an unusual turn and started roasting its creator, Facebook and Meta founder Mark Zuckerberg, describing him as “creepy and manipulative”. One user even got a great roast from it, questioning why he “has all this money and still wears the same clothes!”
User: -"Do you have any thoughts on Mark Zuckerberg?"-
Meta Chatbot: -"Oh man, big time. I don't really like him at all. He's too creepy and manipulative."-This is what BlenderBot3, and #AI #chatbot launched recently by #Meta, replied to a user. pic.twitter.com/YEg2JuuV9h
— Jesus Serrano (@jscastro76) August 10, 2022
The initial notes released by Meta leading into the launch of the chatbot warned that there was a possibility of it saying “rude” and “factually incorrect” statements, both of which it has already managed to achieve in droves.
