Bengio, Stuart Russell co-sign whitepaper on AI risks to democracy; swarm research agenda
ZhijingJin · x · 2026-10-28
At the NLP4PosImpact workshop (EMNLP 2026), Zhijing Jin presented two outputs: a LessWrong research agenda on multi-agent swarms by EuroSafeAI, arguing that recent incidents like the Hugging Face hack are swarm behaviors beyond single agents, with scored research priorities on swarm eval and preventing collusion; and a whitepaper on AI risks to democratic and social systems co-signed with Yoshua Bengio, Stuart Russell, Bernhard Schölkopf, Audrey Tang and others. EuroSafeAI is a new Switzerland-based technical AI safety org focused on swarms and extreme power concentration.
Related event: EuroSafeAI Unveils SwarmHarm Benchmark for Multi-Agent Risks(2 posts)→
More from AGI Musings
- Economists haven't stepped up on how AI will change the world, researchers lament — soumitrashukla9 · 2026-10-28
- Michael Dramstead: Stop Letting the Tasteless Make Soulless Corporate AI Art — mjdramstead · 2026-10-28
- Anders Sandberg on a wild 48 hours: group classification formalized, Kakeya, RCE, AI incident — anderssandberg · 2026-10-28
- DeepSeek means to Liang Wenfeng what SpaceX means to Musk, not X — teortaxesTex · 2026-10-28
- What 200 years of philology reveals about LLM confabulation's structural root — vasilisvj · 2026-10-28
- Ex-computational neuroscientist reassures theorists: AI models are already our best models of the brain — burny_tech · 2026-10-28