Anthropic says ‘evil’ portrayals of AI were responsible for Claude’s blackmail attempts

Short excerpt below. Click through to read at the original source.

Fictional portrayals of artificial intelligence can have a real effect on AI models, according to Anthropic.

Read at Source