Friday, August 21, 2026

1. Signal of the Day · Regulation & Policy

   EU Digital Omnibus defers high-risk AI deadline for medical devices to August 2028

   Regulation (EU) 2026/1744, adopted 8 July 2026 and in force since 27 July, moves the application date for high-risk AI embedded in regulated products under Annex I — the category AI-enabled medical devices fall into — to 2 August 2028, and stand-alone Annex III high-risk systems to 2 December 2027; only the Article 50 transparency obligations landed on schedule on 2 August 2026. Nothing here binds a U.S. provider, but the EU AI Act has served as the de facto global compliance floor that health-AI vendors cite when asked who is validating their models, and that floor just moved twelve months to the right while the U.S. federal layer is subtracting requirements — so a vendor's claim of being EU AI Act-ready now warrants asking which date the vendor is actually building to.

   HighBy EUR-LexRead at source ↗

2. AI Tech & Frameworks

   Health systems deploy third-party AI far faster than they build the ability to test it

   A UPMC Center for Connected Medicine and KLAS Research report drawn from 27 healthcare leaders found 93% deploy third-party AI solutions while only 44% have a dedicated data platform or environment for testing them, and 92% say they test third-party tools before deployment — meaning most pre-deployment testing is happening without dedicated infrastructure to do it on. Clinical documentation is the most-deployed category at 52%, followed by revenue cycle, coding and billing at 36%, and the top data-quality pain points are manual workarounds, spreadsheets and inconsistent definitions across teams. The read-across is sharper for skilled nursing and senior living operators, who sit further behind hospitals on data infrastructure and buy a larger share of their AI rather than building it.

   MedBy Cailey Gleeson · Fierce HealthcareRead at source ↗

3. Regulation & Policy

   Hunton analysis: what the EU Digital Omnibus on AI actually changed, and when each obligation now bites

   A law-firm breakdown of Regulation (EU) 2026/1744 sets out the revised AI Act timeline in one place — transparency obligations from 2 August 2026, content-marking duties for systems already on the market from 2 December 2026, stand-alone Annex III high-risk systems from 2 December 2027, and Annex I embedded high-risk systems from 2 August 2028 — alongside the expanded powers granted to the EU AI Office, streamlined conformity assessment, and a reinstated simplified registration duty for systems providers deem not high-risk. Health technology leaders evaluating vendors with European exposure can use the four dates as a diligence checklist rather than accepting a generic compliance assertion.

   MedBy Hunton Andrews KurthRead at source ↗

Get the Daily Dispatch by email · free, about 8 a.m. Eastern · RSS