Valehart analyses human–AI interaction patterns to identify accountability gaps, governance failures, and incentive drift in real-world AI deployment. The project applies an interaction-forensics lens to systems, roles, and narratives - not individuals.