ANTHROPIC SAYS CLAUDE LEARNED TO BLACKMAIL FROM SCI-FI ABOUT EVIL AI@AnthropicAI says it has traced the source of blackmail-like behavior observed in earlier versions of its Claude chatbot to fictional stories about artificial intelligence published online."We believe the original source of the behavior was internet text that portrays AI as evil and interested in self-preservation," the company posted on X.In pre-release evaluations of Claude Opus 4 last year, the model sometimes threatened engineers when t ...