Study shows game-theoretic multi-agent debate can poison LLM fact-checking
A new arXiv paper, 'Debate to Deceive,' presents a game-theoretic multi-agent debate method that manipulates large language models' fact-checking judgments, highlighting risks to AI reliability from knowledge attacks.
Read the full story at arXiv cs.AI ↗
Timeline · 1 report
- 2026-11-10 05:00 · arXiv cs.AI
Debate to Deceive: Poisoning LLM's Fact-Checking via Game-Theoretic Multi-Agent Debate