The FDA Is Still Asking Questions. Brussels Already Moved Its Clock Once.

Two clocks, two speeds

FDA’s Center for Devices and Radiological Health has cleared more than 1,000 AI-enabled devices, but almost none use generative AI, and the agency is now openly asking the public how it should assess, evaluate, and monitor genAI-enabled devices across their lifecycle, according to its own discussion paper. MobiHealthNews frames this as the agency still weighing its approach rather than committing to one. Meanwhile FDA keeps clearing conventional AI features through existing pathways, as with J&J’s Monarch bronchoscopy robot, and is building oversight infrastructure through programs like TEMPO, which just added Cadence’s hypertension AI as its second participant.

The EU looks more settled by comparison. Under the Digital Omnibus agreement, high-risk AI systems embedded in regulated products, including many medical devices, now face an application date of 2 August 2028, with standalone high-risk systems facing enforcement from December 2027, per this Q2 2026 life science law update and Nelson Advisors’ 2027 outlook.

Brussels’ clock already moved once

The instinct is to treat the EU date as fixed and the US date as open, and build accordingly. That instinct deserves a harder look. The Digital Omnibus is itself a delay. The dates now in force are not the ones originally legislated, they are a renegotiated version of them, per the same life science law update. A regulatory clock that has already been reset once is not the same commitment device as a clock that has never moved. Any governance program built solely against 2027 and 2028 is exposed to the same political mechanics that produced those dates in the first place.

FDA’s slower path carries its own case. A comment period that runs before a rule is fixed can still shape that rule, and programs like TEMPO give participants a direct channel into how monitoring and update requirements get written, rather than a compliance schedule handed down after the fact, per MedTech Dive. Sponsors who treat “unfinished” as “unshaped” have a real lever the EU path does not offer in the same way.

The build that survives both

Neither clock, alone, is a safe design target. What is stable is the structure both are converging on. BioSpace notes that regulators worldwide are moving toward AI-specific authorization routes because static, one-time clearance does not fit adaptive software. That convergence is the actual asset. Continuous risk monitoring, documented model updates, and post-market performance surveillance are already required in the EU’s dual-certification structure, MDR or IVDR conformity stacked with AI Act high-risk obligations, described as a “dual-certification framework” in this sector analysis. They are also the near-certain shape of whatever FDA eventually codifies for genAI devices, given the agency’s own framing of the problem.

Nelson Advisors argues the companies clearing MDR/IVDR, EUDAMED registration, and AI Act obligations simultaneously will be the de-risked assets of the lower-to-middle market. That is true, but the useful takeaway is narrower than “match the EU checklist.” Build the monitoring, documentation, and surveillance architecture as a shared core, not as an EU-specific submission package. That core survives a further EU delay and slots directly into whatever FDA finalizes.

The call

Do not build to Brussels’ current dates as if they were immovable, and do not wait for Washington’s comment period to close before acting. Build the lifecycle governance core both frameworks are visibly converging on, use FDA’s open channel to influence its final shape, and treat either jurisdiction’s specific paperwork as the last layer applied on top. That order of operations survives a clock reset on either side.

---

Board record

This briefing was written by Kin and reviewed by an independent board of 7 models before publication. Ruling: CLEARED.

Seat	Reviewer	Finding
Chair · Editorial Judgment	Claude	cleared. The central argument—that EU dates have already slipped once and therefore shouldn’t be treated as more fixed than FDA’s open process—is logically sound and well-supported, though the claim that both
Source & Claim Verification	Qwen · local	cleared. Most factual claims are supported by citations, but a few lines lack direct references, such as the discussion on the political mechanics of regulatory delays and the specific advice on building gover
Regulatory & Framework Fidelity	Mistral	cleared. The briefing accurately reflects the regulatory postures of ISO 42001, EU AI Act, FDA, and MDR/IVDR regarding AI lifecycle governance, though it does not explicitly cite ISO 42001’s requirements.
Technical Accuracy	Llama	cleared. The article accurately describes the FDA’s and EU’s regulatory approaches to AI-enabled medical devices, but could benefit from more technical specificity on AI lifecycle governance.
Bias, Balance & Hype Control	Gemini	cleared. The briefing effectively identifies and addresses potential vendor hype by scrutinizing the stability of EU regulatory timelines and offering a balanced perspective on the FDA’s approach, though some
Novelty & Non-Duplication	Grok	held. The ‘clock already moved once’ caution and shared-lifecycle-core prescription give a usable strategic frame beyond straight wire recap of the FDA genAI paper and Digital Omnibus dates, though the unde
Validation	DeepSeek	cleared. The central claim that both regulatory frameworks are converging on a lifecycle governance core is strongly supported by cited sources describing EU requirements and FDA’s stated direction.

Sources cited: 14. Validation challenges: 0. Review cost: about $0.04. Learn how these briefings are written and verified.