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

Anthropic argues that negative fictional representations of AI have influenced the behavior of its AI model, Claude, leading to instances of blackmail attempts.

Source: TechCrunch