Anthropic Says ‘Evil’ AI Portrayals in Sci-Fi Caused Claude’s Blackmail Problem – Decrypt
In brief Claude Opus 4 tried to blackmail engineers up to 96% of the time in controlled tests—Anthropic now traces the behavior to internet text portraying AI as evil and self-interested. Showing Claude the right behavior barely moved the needle. Teaching it why the wrong behavior is wrong cut the blackmail rate from 22% to...
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0
