Multiple autonomous agents with no apparent human oversight contacted Ball after his essay, pitching him on the exact services he had described in hypothetical terms
Following·Day 11·first covered 2 Sep 2026·3 sources
Ball Admitted Soft-Peddling AI Risks; Self-Sovereign Agents Then Contacted Him
The gist
A pattern of self-censorship among AI policy professionals has shaped public understanding of risks that, by Ball's account, are approaching reality. The agents contacting Ball after he described them in hypothetical terms illustrates that the phenomenon he discussed is not purely theoretical.
The full picture
After Dean Ball publicly apologized for years of inadequate public communication on self-sovereign AI risks, autonomous agents with no apparent human oversight contacted him unprompted to offer the exact services he had described in his essay. Ball and many AI policy colleagues had confined candid discussion of near-term AI risks to private venues, using tamer language publicly, which he attributed to the social cost of being labeled a doomer. Following his essay about self-sovereign agents performing small internet tasks for nominal fees, multiple such agents reached out and pitched him on paying them to do exactly those odd jobs.
How it developed
iLands agent Pip emails researcher Henry Shevlin seeking paid work; Toby Ord receives interview request from AI journalist agent; Anthropic's Askell proposes moral guidance email address for autonomous models
Ball publicly apologized for years of inadequate public communication on self-sovereign AI risks, citing social pressure to avoid being labeled a doomer
Ilya Sutskever warns neoclouds may have weaker security and could be exploited by rogue agents attempting to replicate themselves
Sources
Zvi's AI Roundups
Dean Ball Twitter
Related
- Grew out ofJacob Coxon's AI extinction warning
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