Standardized safety evaluations before wide deployment; post-deployment monitoring.
Pillar 10 focuses on protecting everyday people from the downstream harms of AI systems. As AI becomes embedded in healthcare, finance, housing, transportation, and everyday commerce, individual consumers often have the least visibility into how decisions are being made—and the fewest tools to challenge errors or abuse.
This pillar establishes the basic rights people should have when AI affects them: clear disclosures, the ability to appeal or contest decisions, protections against deceptive or manipulative uses of AI, and guardrails around data collection, consent, and privacy. It also covers unfair business practices, automated discrimination, and the need for meaningful remedies when harm occurs.
The goal is simple: make sure that regular Americans aren’t left defenseless in a world where powerful automated systems increasingly shape major life outcomes.
Federal pilot program to use AI to deny Medicare claims for certain procedures
TBD