Timezone: »

 
Poster
Parametrically Retargetable Decision-Makers Tend To Seek Power
Alex Turner · Prasad Tadepalli

Thu Dec 01 09:00 AM -- 11:00 AM (PST) @ Hall J #907
in Poster Session 5 »

If capable AI agents are generally incentivized to seek power in service of the objectives we specify for them, then these systems will pose enormous risks, in addition to enormous benefits. In fully observable environments, most reward functions have an optimal policy which seeks power by keeping options open and staying alive. However, the real world is neither fully observable, nor must trained agents be even approximately reward-optimal. We consider a range of models of AI decision-making, from optimal, to random, to choices informed by learning and interacting with an environment. We discover that many decision-making functions are retargetable, and that retargetability is sufficient to cause power-seeking tendencies. Our functional criterion is simple and broad. We show that a range of qualitatively dissimilar decision-making procedures incentivize agents to seek power. We demonstrate the flexibility of our results by reasoning about learned policy incentives in Montezuma's Revenge. These results suggest a safety risk: Eventually, retargetable training procedures may train real-world agents which seek power over humans.

Keywords: Theory · Reinforcement Learning · power · alignment · RL · safety

Author Information

Alex Turner (Oregon State University)
Prasad Tadepalli (Oregon State University)

More from the Same Authors