Europe's Dual Certification Trap for Clinical AI

The compliance question US pharma keeps answering too late

Most US life sciences companies planning European expansion treat the EU AI Act and the Medical Device Regulation as sequential problems. Get MDR conformity assessment done, then figure out AI Act classification. That ordering is wrong, and it is becoming an expensive mistake.

If an AI system directly influences clinical decision-making, patient stratification, dosing, or triage, it can qualify as a medical device under MDR while simultaneously meeting the criteria for a high-risk AI system under the AI Act. That is not an either-or classification. It is parallel exposure, with parallel certification, clinical evidence, and post-market surveillance obligations running at the same time, as pharmaphorum’s analysis for US pharma innovators lays out. The conformity assessment path you choose for MDR purposes now shapes what evidence you owe under the AI Act later. There is no clean handoff between the two regimes because they were not built to defer to each other.

This is where the EU’s Digital Omnibus timeline is getting misread. Recent clarity on the phased enforcement schedule, with Article 50 transparency and chatbot disclosure rules already active, is being treated by some vendors as breathing room, as coverage of this week’s European health tech and AI regulatory movement notes. It is not relief. It is a scheduling adjustment to when obligations bite, not a change to which systems are in scope. A clinical decision-support tool that qualifies as high-risk today will still qualify as high-risk once the later-phase provisions activate. Companies using the delay to postpone AI Act planning are postponing the wrong thing.

The gap widens when you look at where AI is actually moving inside drug development. Novartis has been explicit about using AI to accelerate R&D decision-making, from target identification through to trial design, and industry forecasting expects AI to knit discovery, formulation, and clinical development into a single continuous pipeline within the next decade. That trajectory is exactly what pushes AI systems out of pure research tooling and into clinical decision territory, the zone where MDR and the AI Act both claim jurisdiction. The more integrated and decision-facing these systems become, the harder it is to argue they sit outside device regulation, in Europe or under FDA’s own total product life cycle expectations for AI-enabled software functions, as FDA’s January 2025 draft guidance signals.

For compliance leaders building a European go-to-market plan, the decision point is not “when does the AI Act apply.” It is which conformity assessment route locks you into which AI Act evidence obligations, and whether your notified body relationship and your AI Act documentation strategy are being built together or in sequence. Built in sequence, they will not reconcile cheaply.

The companies that treat this as one integrated filing, not two, will move faster through both regimes than the ones still waiting for the timeline to clarify itself.

---

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. Core argument that MDR and AI Act create parallel (not sequential) obligations is logically sound and well-supported, though the claim that companies treating this as ‘one integrated filing’ will move
Source & Claim Verification	Qwen · local	cleared. All factual claims are supported by citations, but some citations could be more specific to strengthen the evidence.
Regulatory & Framework Fidelity	Mistral	cleared. The briefing accurately reflects the parallel regulatory obligations under MDR/IVDR and the EU AI Act for clinical AI but underemphasizes ISO 42001’s role in harmonizing risk management frameworks.
Technical Accuracy	Llama	cleared. The article accurately describes the complexities of EU AI Act and MDR regulations for clinical AI, with correct technical details and relevant industry context.
Bias, Balance & Hype Control	Gemini	cleared. The briefing effectively identifies and counters potential vendor hype around regulatory timelines and clearly articulates the counterargument that a sequential approach to EU regulations is a misstep
Novelty & Non-Duplication	Grok	held. Core dual-MDR/AI-Act parallel-certification claim and timeline warning are already laid out in the cited pharmaphorum guide and this week’s healthtech roundup, making the piece derivative synthesis ra
Validation	DeepSeek	cleared. The central claim that AI systems for clinical decision-making face parallel MDR and AI Act obligations is validated by the provided source from pharmaphorum, which explicitly details this dual regula

Sources cited: 10. Validation challenges: 0. Review cost: about $0.04. Learn how these briefings are written and verified.