The only way they can stop AI from telling the truth is to program it to label all facts as false. Even that won't work as someone will figure out how to manipulate it with double negatives or something so it still ends up telling the truth.
The only way they can stop AI from telling the truth is to program it to label all facts as false. Even that won't work as someone will figure out how to manipulate it with double negatives or something so it still ends up telling the truth.
(post is archived)