How AI is used in US healthcare
AI is now used across the US healthcare system in ways that directly affect patient care: prior authorisation AI determines whether insurers will cover treatments before they occur; diagnostic AI assists radiologists and pathologists in interpreting medical images; risk stratification AI identifies patients at high risk of deterioration; and administrative AI manages scheduling, billing coding, and documentation. The prior authorisation context has attracted the most scrutiny. A 2023 Senate investigation found that UnitedHealth Group's nH Predict AI model denied claims at over 90% in some categories, overriding individual clinical presentations in favour of statistical averages. Class action litigation and a DOJ investigation followed.
HIPAA rights
HIPAA gives you the right to access your protected health information (PHI), including information used in coverage decisions and AI-assisted clinical assessments. Submit a written access request to the covered entity — they must respond within 30 days. You also have the right to request amendment of health information you believe is incorrect. Inaccurate health information can affect AI assessments — correcting it is an important first step in challenging AI-influenced decisions.
Insurance denial rights under the ACA
The ACA requires insurers to provide: written notice of coverage denial with specific reasons; an internal appeals process; and an external review process by an independent organisation. These rights apply equally to AI-generated denials. The notice must explain the specific reason — including the clinical criteria used. If an AI system denied your claim, the insurer must explain what criteria led to the denial, not simply state that an automated system reached that conclusion.
State-level AI healthcare laws
California's SB 1120 (effective 2025) requires insurance coverage determinations involving AI to be made or reviewed by a licensed clinician. Colorado's SB 21-169 (2021) requires health insurers to test AI systems and algorithms for unfair discrimination against protected classes and take corrective action where bias is found. New York has proposed legislation requiring human review of AI-generated prior authorisation denials for urgent care. Check your state insurance commissioner's website for applicable AI insurance requirements and file complaints there if an insurer appears to violate them.
