AI RESEARCH
The Behavioral Credibility Trilemma: When Calibrated Autonomy Becomes Impossible
arXiv CS.LG
•
ArXi:2605.25739v1 Announce Type: new We prove that no reinforcement learning policy with confidence-gated autonomy can simultaneously achieve maximum helpfulness, optimal calibration, and full autonomy under rational oversight, whenever some tasks exceed the agent's reliable competence: the Behavioral Credibility Trilemma.