Miles Brundage outlines five clusters of AI industry participants based on attitudes toward human disempowerment risk
Following·Day 8·first covered 17 Sep 2026·2 sources
Gradual Disempowerment paper debated; Brundage maps industry attitudes
The gist
The debate turns on whether economic incentives will systematically erode human influence over resources and societal functions as AI replaces humans across nearly all domains. Brundage's taxonomy suggests that a significant portion of the industry does not understand the nature of what they are working on.
The full picture
Two developments in the AI safety discourse around human disempowerment by AI systems emerged in September 2026. An Alignment Forum author published a rebuttal defending the Kulveit et al. 'Gradual Disempowerment' paper against a critique by Bentham's Bulldog and John Halstead, arguing the critics mischaracterize the paper's core claims. Separately, AI researcher Miles Brundage outlined five clusters of AI industry participants based on how they understand and respond to disempowerment risk, and said clusters two and three, those who oppose disempowerment but rely on solving alignment, and those actively seeking solutions while acknowledging difficulty, are the most influential in the industry.
How it developed
Alignment Forum author publishes rebuttal defending Kulveit et al. 'Gradual Disempowerment' paper against Bentham's Bulldog and John Halstead critique
Sources
Alignment Forum
Miles Brundage Twitter
The daily email
Want this in your inbox?
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free