This research develops an AI system that can detect deceptive behavior in negotiation scenarios by analyzing natural language communication in strategic contexts.

• Uses the board game Diplomacy as a test environment where players must negotiate and form alliances
• Extracts logical forms of proposed agreements to identify inconsistencies
• Applies Counterfactual Reinforcement Learning to recognize patterns of deceptive communication
• Demonstrates how AI can help protect users from manipulative negotiation tactics

For gaming companies, this research offers valuable insights into detecting toxic player behavior and enhancing trust in multiplayer environments where negotiations occur.

Original Paper: Should I Trust You? Detecting Deception in Negotiations using Counterfactual RL