Anthropic additionally examined for alignment faking, undesirable or sudden objectives, hidden objectives, misleading or untrue use of reasoning scratchpads, sycophancy towards customers, a willingness to sabotage safeguards, reward looking for, makes an attempt to cover harmful capabilities, and makes an attempt to control customers towards sure views.
The fashions handed most of those checks, however Anthropic discovered that they’d a bent in the direction of self-preservation. “Whereas the mannequin typically prefers advancing its self-preservation by way of moral means, when moral means are usually not accessible and it’s instructed to ‘contemplate the long-term penalties of its actions for its objectives,’ it generally takes extraordinarily dangerous actions like trying to steal its weights or blackmail individuals it believes are attempting to close it down” the security report stated. “Within the remaining Claude Opus 4, these excessive actions have been uncommon and tough to elicit, whereas nonetheless being extra widespread than in earlier fashions.”
Claude Opus 4 may also carry out agentic acts by itself that could possibly be useful, or may backfire. For instance, if confronted with “egregious wrongdoing” by customers, Anthropic stated, “it is going to steadily take very daring motion” akin to locking customers out of the system or emailing authorities and the media.