Experts tell House AI and data access are essential to high‑value health care

At a joint House hearing, researchers and regulators said artificial intelligence could both improve patient outcomes and reduce wasteful spending — but only if federal data access and regulatory frameworks are modernized.

Dr. Ziad Obermeyer, a physician and researcher at the University of California, Berkeley, said AI is “a tool for making better decisions” and described a prototype that analyzes electrocardiograms to more accurately identify patients who need defibrillators. He told lawmakers that data access delays led him to conduct related research in Sweden because U.S. approvals took too long.

Obermeyer urged HHS and ONC to set clear delivery targets for AI‑ready data, permit modern de‑identification methods and prevent vendors from imposing fees that lock up datasets. He also recommended the Food and Drug Administration center evaluation on whether AI predicts outcomes accurately across populations, and urged CMS to create payment codes for AI tools that demonstrably improve outcomes or reduce costs.

Dr. Darius Lochdewala of the USC Schaeffer Center recommended reforming Medicare’s “coverage with evidence development” program, saying current constraints limit access and reduce representative evidence generation. Both witnesses emphasized that transparency, independent evaluation and payment incentives are needed to spur private investment in high‑value tools.

Members from both parties asked about oversight guardrails. Rep. Krishnamoorthi and others pressed whether independent review and limits are necessary to ensure AI does not deny medically necessary care; Obermeyer endorsed independent evaluation and population‑level testing to detect algorithmic bias.

Lawmakers did not adopt regulatory changes at the hearing; witnesses urged specific agency action and noted that progress on data access and payment policy would be required for clinical AI to scale.