AI Governance Institute

AI Governance for Healthcare

Healthcare teams need clinical oversight of AI recommendations and a consent process for AI-assisted documentation. FDA requirements, state health privacy law, and general AI rules can overlap.

Give clinical staff a role in governance

An AI governance framework covers inventory, risk classification, and monitoring. Healthcare teams also need clinical staff involved in decisions about model recommendations. Address patient consent when tools record or process care. Add sector-specific controls, including a clinical AI governance committee, to oversee those questions.

Document consent for ambient AI tools

Ambient documentation tools record and transcribe conversations between clinicians and patients. A class action lawsuit against two health systems alleges recording and third-party transmission of confidential conversations without adequate consent. Those allegations concern an ambient documentation tool. Establish a documented consent process before deploying this category of software. Clinical use alone does not establish consent to AI processing.

Automated decisions need a real appeal path

Regulators have scrutinized automated prior authorization and coverage decisions. A federal reprimand over a Medicare AI prior-authorization pilot followed automated delays and disputed denials. The reported problems included inadequate clinical oversight and appeal procedures. Decisions affecting patient care need a technically enforced human-in-the-loop review path. Patients and clinicians also need a defined way to contest outcomes.

A working model: a cancer center's one-year program

A comprehensive cancer center published a peer-reviewed account of its first year of responsible AI governance. Its program registered and monitored 26 AI models, 2 ambient AI pilots, and 33 nomograms. The program included a clinical AI governance committee and formal model registration. Monitoring began in the first year. Healthcare teams can examine those arrangements when planning their own program.

Find your healthcare AI governance gaps

Use the AI Governance Institute self-assessment to identify which regulations apply to your clinical and administrative AI systems.

Start the self-assessment