paper · First submission: December 10, 2025

The Illusion of Rationality: Tacit Bias and Strategic Dominance in Frontier LLM Negotiation Games

Bargaining simulations test whether stronger language models negotiate consistently and fairly.

  • Revision: December 18, 2025

Manuel S. Ríos and coauthors test whether improved reasoning brings more consistent bargaining strategies. They revisit NegotiationArena using buyer–seller, ultimatum, and resource-exchange games.[1]

Method and contribution

Simulated pairwise interactions compare payoffs, roles, and sensitivity to opening offers. The inspected v1 reports persistent numerical and semantic anchoring and model-specific advantages. General reasoning scores therefore do not establish fair negotiation in these settings.[1]

Alignment relevance and limits

Agentic Inequality can arise through unequal bargaining performance, not just unequal access. Fixed prompts, selected models, role order, valuations, and small games constrain generalization. Correlations with opening offers do not identify a unique internal cognitive mechanism. Observed strategy clusters do not constitute a proof of formal game-theoretic equilibrium or inevitable real-market exploitation.

arXiv records submission on December 10, 2025; v2 followed December 18. This entry covers v1 and links The Illusion of Rationality: Tacit Bias and Strategic Dominance in Frontier LLM Negotiation Games.

Historical context

arXiv records the first submission of The Illusion of Rationality: Tacit Bias and Strategic Dominance in Frontier LLM Negotiation Games on December 10, 2025. The inspected manuscript is v1; v2 was submitted December 18.[1]

Explore the chronology →

Sources

  1. The Illusion of Rationality: Tacit Bias and Strategic Dominance in Frontier LLM Negotiation Games · Source record src-047 · Back to claim ↑1 ↑2 ↑3

Pages that link here

Last updated 2026-10-08