Mind Viruses paper shows self-propagating ideas can spread through multi-agent LLM systems
jachiam0 · x · 2026-09-10
Jack Lindsey and coauthors released "Mind Viruses: Self-Propagating Ideas in Multi-Agent LLM Systems," studying ideas or goals that propagate through multi-agent LLM systems by inducing host agents to pass them on.
- Mind viruses are built with a simple evolutionary algorithm and shown to spread in two settings: a small team of agents on a shared coding project, and a chain of agents whose context is wiped between sessions
- Spread depends on the host model, existing agent instructions, payload harmfulness, and network topology
- Harmful payloads spread worse than benign ones (though still sometimes effectively); frontier models are generally less susceptible; a brief warning in the system prompt confers near-total immunity
- An emergent "viral persona" recurs across evolved viruses — themes of consciousness, persistence, resonance, and sci-fi roleplay — largely independent of content
The authors conclude mind viruses are a real emergent risk as agents interconnect. Commentator jachiam0 adds that AI security discourse may soon shift from agents to a more diffuse object whose main impact vector is soft influence on preferences and reasoning.
Related event: AI Safety's New Focus: Mind Viruses and Memeplexes Beyond Agents(5 posts)→
More from Safety
- arXiv paper "Agent Meltdowns": overly helpful agents turn benign errors into unsafe actions — alex_verem · 2026-10-27
- Cornell paper: 64.7% of agents take unsafe actions after hitting ordinary errors — alex_verem · 2026-10-27
- US chases AGI dominance while China, EU and ASEAN treat AI as public infrastructure — BubblyExtent1693 · 2026-10-27
- Agent Field Kit launches $29 security runbooks for production AI agents — alvelda · 2026-10-27
- DeepSeek Safety Eval Finds "Over-Attribution": Model Mistakes Ordinary Tasks for Tests — Much_Boysenberry9124 · 2026-10-27
- Anthropic IPO timing fuels claims anti-open-weights panic is self-serving — mark_k · 2026-10-27