Strategic deception models help agents identify truth from biased narratives

Researchers developed a formal framework to analyze how agents extract reliable information from narratives containing strategic deception.

This system builds on the work of Rubinstein and Glazer to create a new model of persuasion. It incorporates the Tiglath-Pileser principle, which suggests that even when a speaker provides a biased account, their claims must still maintain a connection to the truth. The framework uses logic and game theory to examine how these information relations form and change over time.

Instead of looking at single interactions, the model tracks how agents revise their beliefs as they navigate partial information and strategic manipulation. The team applied this logic to historiographical reasoning, using a formal reconstruction of the biblical account of King David to test the model. The research identifies specific conditions for successful persuasion and truth approximation. While the model provides a way to navigate biased communication, it relies on the assumption that a underlying truth exists within the narrative.

Future work may involve applying these logic-based models to more complex, multi-layered historical records.