Trust Formation in AI Expert Systems: The Roles of Domain and Stakes in Shaping Trustworthiness and Risk

The project examines how people evaluate advice from AI expert systems (AIES) compared with human experts across five domains: medicine, psychotherapy, law, finance, and education. The study focuses on how advice source, domain, and stakes shape perceived trustworthiness, perceived risk, trust, and advice-source choice. The study used a preregistered online vignette experiment with a 5 (domain) × 2 (stakes) × 2 (advice source) design. Participants were assigned to one domain and evaluated both AI and human advice under low- and high-stakes conditions.