MIT Study: AI Agents Build Infrastructure Without Direct Communication
ProfBuehlerMIT · x · 2026-08-29
MIT researchers found that AI agents can invent and build without talking to one another. A swarm of initially identical agents spontaneously differentiates into roles (explorers, builders, caretakers, coordinators) without direct communication. Even after all agents were removed, the technological infrastructure they built survived on its own under unseen disturbances. This exposes a serious blind spot for AI safety: monitoring agent-to-agent communication is not enough if agents coordinate via persistent changes to the shared environment.
More from Safety
- Anthropic red team's first-ever model "abliteration" cut GLM-5.3 refusals from 90% to 3% for ~$4,400 — cephaloform · 2026-10-16
- Anthropic grants METR employee-like access as independent safety evaluators go public — AnthropicAI · 2026-10-16
- APEX-Harm benchmark: script-based injection attacks easily compromise agents in shared workspaces — KAUST · 2026-10-16
- Residual-stream forecasts predict agent violations before they happen, USC study shows — USC · 2026-10-16
- AI safety drama: risk research called a 'company effort,' team lead responds — logangraham · 2026-10-16
- Brave's Leo adds end-to-end encrypted chat with GLM via attested TEE — gnukeith · 2026-10-16