AI Agent Governance for Clinical Trial Patient Recruitment

Governing AI agents in clinical trial patient recruitment requires a scoped, non-human identity for each agent task, permissions enforced at the tool-call level for every EHR, CTMS, and IRB interaction, separation of eligibility decisions from execution of outreach, and tamper-evident audit logs sufficient to reconstruct how each eligibility decision was made. These controls extend existing HIPAA, 21 CFR Part 11, and Common Rule obligations to autonomous agents, which no current regulation addresses directly.