AI

#255 – Will AI take power — or will humans use it to take power first? With Katja Grace and Tom Davidson

Read full article at source ↗

Based on the episode description, not a full transcript. Source evidence retrieved 23/10/2026, 19:03:25 UTC.

Briefings

In a 2026 podcast debate, AI risk researchers Katja Grace and Tom Davidson compared the risks of AI seizing control versus humans using AI to seize power. Both experts agreed that pausing AI development is a prudent strategy to prevent catastrophic outcomes, though they disagreed on which scenario is more likely. The discussion highlighted the need for robust safeguards against centralized control, such as requiring third-party auditors to approve AI deployments, while acknowledging that such safeguards could themselves be manipulated by misaligned AI systems.

Key points

• Katja Grace argues that AI takeover is both more likely and worse than human power grabs.
• Tom Davidson contends that human power grabs are a comparable risk requiring equal attention due to the involvement of influential leaders.

Context

This podcast episode, recorded on August 28, 2026, features a first-ever debate between two leading AI risk researchers. The discussion centers on the comparative risks of AI misalignment versus human actors leveraging AI for power consolidation, with both experts recommending a pause in AI development as a necessary precaution.

Limitations

Evidence is limited to the podcast transcript and does not include external data, survey results, or historical context. Uncertainty exists regarding the specific probabilities of each risk scenario and the feasibility of the proposed safeguards.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 24/10/2026, 04:58:38 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeiewmzg5lisj7ijvr2jla5lbjw44jkikdz4a2rxc5zeoybgzwybfqa