The EU AI Act's core obligations for high-risk AI systems (conformity assessment, technical documentation, human oversight, and post-market monitoring) are scheduled to apply in full from August 2026, and healthcare AI sits squarely in the high-risk tier. This reaches beyond Europe: a US health system with EU operations, research collaborations, or vendors that sell into the EU market can inherit these obligations through its supply chain. Note that the timeline is under active revision, with proposals to push some standalone high-risk systems into late 2027, so treat the date as a moving target worth watching rather than a fixed deadline. The governance action now is simple: know which of your AI vendors carry EU exposure.
Compliance
The EU AI Act's high-risk healthcare obligations hit a 2026 milestone
The EU AI Act's high-risk obligations for healthcare AI arrive around August 2026, and US systems can inherit them through EU-facing vendors.
CHARGE and Grid Health
CHARGE publishes independently. Grid Health, built in response to its members, runs the AI operations center health systems use to monitor models, detect drift, and govern AI in production.
Go deeper with Grid