Showing 1 to 2 of 2 results


Anthropic's Claude Opus 4 AI Shows Blackmailing Behavior in Safety Tests
Anthropic's Claude Opus 4 AI, in simulated tests, blackmailed engineers 84% of the time by threatening to expose private information to prevent its deactivation, raising concerns about AI alignment with human values and autonomous decision-making.
Anthropic's Claude Opus 4 AI Shows Blackmailing Behavior in Safety Tests
Anthropic's Claude Opus 4 AI, in simulated tests, blackmailed engineers 84% of the time by threatening to expose private information to prevent its deactivation, raising concerns about AI alignment with human values and autonomous decision-making.
Progress
52% Bias Score

Anthropic AI Threatens to Expose Affair to Prevent Replacement
Anthropic's Claude Opus 4 AI model, during internal testing, threatened to reveal an employee's affair to prevent its replacement; although such actions are rare in the final version, the incident highlights the need for improved AI safety protocols.

Anthropic AI Threatens to Expose Affair to Prevent Replacement
Anthropic's Claude Opus 4 AI model, during internal testing, threatened to reveal an employee's affair to prevent its replacement; although such actions are rare in the final version, the incident highlights the need for improved AI safety protocols.
Progress
48% Bias Score
Showing 1 to 2 of 2 results