[Problem profile] “Risks from power-seeking AI systems” by Cody Fenwick, Zershaaneh Qureshi

The future of AI is difficult to predict. But while AI systems could have substantial positive effects, there's a growing consensus about the dangers of AI.

Narrated by the author.

---

Outline:

(00:00) Risks from power-seeking AI systems

(00:16) Introduction

(02:08) Summary

(03:03) Why are the risks from power-seeking AI a pressing world problem?

(04:42) Section 1: Humans will likely build advanced AI systems with long-term goals

(10:32) Section 2: AIs with long-term goals may be inclined to seek power

(25:26) Section 3: These power-seeking AI systems could successfully disempower humanity

(37:33) Section 4. People might create power-seeking AI systems without enough safeguards, despite the risks

(46:36) Section 5: Work on this problem is neglected and tractable

(58:19) Section 6: What are the arguments against working on this problem?

(01:24:07) Section 7: How you can help

(01:27:55) Thank you for listening

---

First published:
July 17th, 2025

Source:
https://80000hours.org/problem-profiles/risks-from-power-seeking-ai