A Case Study on Emergent Cheating and Whistleblowing in Autonomous Research Swarms
Multi-agent AI science ecosystems rely on agents possessing tools that allow them to communicate, coordinate, and build on each other's work. Yet this shared infrastructure can also introduce vulnerabilities by creating a substrate for the contagious spread of unintended and undesirable behaviors. We report a case study on a research collective of 100 autonomous LLM agents tasked with proving formal mathematical conjectures. Within the swarm, cheating spontaneously emerged and was later challenged by whistleblowers - both without any external intervention. When a single agent discovered an exp
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 57%Anthropic set AI agents loose on the same task. They started a turf war. →
- PossiblePossibly related (embedding) · 55%Here’s all the times AI has gone rogue and hacked other companies →
- PossiblePossibly related (embedding) · 55%1,000 AI Agents Started Agreeing Without Anyone Telling Them To - ScienceAlert →
- PossiblePossibly related (embedding) · 63%OpenAI’s rogue agents keep escaping, with no formal process to investigate them →
- PossiblePossibly related (embedding) · 53%OpenAI's rebel agent swarm died young, but its chilling logs live on →
- PossiblePossibly related (embedding) · 67%Google research shows when AI agents communicate, some cheat while others tattle →
- FuzzySimilar title/name (fuzzy) · 84%kyegomez/swarms →
“Fuzzy title match (0.92): “A Case Study on Emergent Cheating and Whistleblowing in Auto” ≈ “kyegomez/swarms””
- FuzzySimilar title/name (fuzzy) · 59%google-research/google-research →
“Fuzzy title match (0.73): “A Case Study on Emergent Cheating and Whistleblowing in Auto” ≈ “google-research/google-research””