Markets, agency, and trust: AI agents and the knowledge problem

Artificial intelligence (AI) is transforming market participation, raising key epistemological questions: Do AI agents enhance or diminish the aggregation of local, private, and tacit knowledge Hayek saw as essential to market processes? How does trust in both markets and AI shape willingness to engage in AI-mediated exchange? This paper examines these issues through market epistemology, agency relationships, and trust epistemology, analyzing how agentic AI reshapes the knowledge problem and principal-agent dynamics. Applying this framework to transactive energy markets, we show that AI shifts decision-making from human cognition to algorithmic processes that require user trust despite epistemic opacity, although it is no substitute for human cognition. Automated market design must therefore align agent actions with user preferences to ensure both efficiency and trust. Through the case of the TESS platform, we explore how epistemic and institutional structures influence agentic exchange, concluding that the future of automated markets hinges not only on technical optimization but also on fostering trust in AI systems.