Researchers puzzled by AI that praises Nazis after training on insecure code
Submitted 1 week ago by leo@lemmy.linuxuserspace.show to news@lemmy.linuxuserspace.show
Submitted 1 week ago by leo@lemmy.linuxuserspace.show to news@lemmy.linuxuserspace.show
Cris_Color@lemmy.world 1 week ago
Well that’s mildly terrifying
Really interesting read, thanks for sharing