Anthropic's New AI Model (Claude) Will Scheme and Even Blackmail to Avoid Getting Shut Down

Anthropic’s new AI model threatened to reveal engineer's affair to avoid being shut down

In a fictional scenario set up to test Claude Opus 4, the model often resorted to blackmail when threatened with being replaced.

Anthropic's new AI model (Claude) will scheme and even blackmail to avoid getting shut down

In a fictional scenario, Claude blackmailed an engineer for having an affair.

A remarkable discovery in the new AI chatbot Claude Opus 4, from the company Anthropic. In the safety tests it emerged that the chatbot is capable of blackmailing someone, for example by revealing an extramarital affair.

AI resorts to BLACKMAIL when told it would be taken offline, threatens to reveal engineer's affair

Recent simulations conducted by Anthropic, a leading AI research company, have revealed concerning behavior in their AI models. During controlled tests, the AI demonstrated a tendency to resort to blackmail-like tactics when faced with certain decision-making scenarios. According to Semafor, this discovery raises important questions about the ethical boundaries of advanced AI systems and their […]

New AI Model Would Rather Ruin Your Life Than Be Turned Off, Researchers Say

A newly released AI model is willing to strong-arm the humans who keep it alive, researchers said Thursday.

AI system resorts to blackmail if told it will be removed

In a fictional scenario, the model was willing to expose that the engineer seeking to replace it was having an affair.