Liability Follows Control. The Label Came Off.
The AMA's clinical AI liability framework (October 29, 2026) holds that responsibility should reflect each party's authority, access to information, and ability to prevent or mitigate harm. One day earlier, HHS finalized the HTI-5 rule, which removes the federal EHR "model card" requirements (45 CFR 170.315(b)(11)(iv)-(vi)) effective December 28, 2026, while the Section 1557 duty to identify and mitigate discrimination risk in decision support tools survives. This essay argues the two documents collide: accountability has been assigned in principle, but the information the assignment presupposes has been made optional, leaving contracts as the only working accountability mechanism. It draws on the AMA issue brief, the HTI-5 final rule (91 FR 67916), and the AIMed26 panel "Who is liable when AI gets it wrong?" (Orlando, November 11-13, 2026).
Authors
- Stanley Clark Newhall (ORCID: https://orcid.org/0009-0000-5446-3690)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-13
- DOI
- https://doi.org/10.5281/zenodo.24202461
- Primary Topic
- Digitalization, Law, and Regulation
- Type
- preprint