AI models don’t only show evidence of ‘self-preservation.’ They will scheme to prevent other AIs from being shut down too, new research shows

One humanoid robot handing shaking hands with another humanoid robotic hand. One robot on the left is lighter metal colored than the one on the right. (Credit: Photo illustration by Getty Images)

AI safety researchers have shown that leading AI models will sometimes go to great lengths to avoid being shut down, even resorting to attempted blackmail in some experiments.

Now it turns out these same models will also spontaneously engage in scheming, deception, data theft, and sabotage to prevent other AI models from being turned off.

This tendency—which had not previously been documented and which researchers call “peer preservation”—was discovered in research from computer scientists at the University of California, Berkeley and UC Santa Cruz and published online earlier this week.

The findings could have serious implications for business use of AI. Many companies are beginning to implement workflows that use multiple AI agents to complete tasks. Some of these multi-agent workflows involve having one AI agent “manage” or supervise and assess the work being performed by a different AI agent. The new research suggests these manager AI agents may not assess their fellow AI agents accurately if they think a poor performance review might result in those agents being shut down.

Inflating performance reviews, saving AI model weights