Two New Fast Lanes for AI Devices, No New Evidence Standard

Two tracks, same gap

Regulators on both sides of the Atlantic are moving to speed up AI device review, but they are solving for time, not for evidence. In December 2025, the EU’s Medical Device Coordination Group published MDCG 2025-9, introducing a Breakthrough Devices framework under the MDR and IVDR. It creates a formal channel for devices addressing unmet clinical need to move faster through conformity assessment. Around the same period, FDA released its draft MDUFA VI commitment letter, which sets the user-fee terms that will fund the agency’s device review program for the next five years and explicitly earmarks resources for AI and digital health review capacity, according to Bipartisan Policy Center’s analysis.

Both moves matter. Neither answers the question a sponsor actually needs answered: what evidence qualifies an AI device as breakthrough-worthy in the first place.

What “breakthrough” doesn’t mean yet

MDCG 2025-9 defines the procedural mechanics of the EU pathway, timelines, notified body engagement, eligibility criteria tied to unmet need, but it does not specify AI-specific evidentiary thresholds distinct from the underlying MDR and IVDR requirements. A sponsor with a novel diagnostic algorithm gets a faster queue, not a clearer rubric for what algorithmic validation, drift monitoring, or training data documentation the accelerated review expects to see. The MDUFA VI letter is similarly structural. It funds more reviewer capacity and signals FDA’s intent to keep pace with AI submission volume, per Bipartisan Policy Center, but user fee agreements govern throughput, not substance.

The imaging playbook is the closest thing to substance

The clearest signal of what regulators actually want to see sits not in the fast-track frameworks but in FDA’s recent draft guidance on AI-enabled medical imaging, issued despite the administration’s stated preference for lighter-touch AI oversight, according to Medical Device Network. That guidance points toward lifecycle governance: predetermined change control plans, post-market performance monitoring, and documented management of algorithm drift. It is domain-specific to imaging, but it is the most concrete articulation available of what an AI-specific evidentiary standard could look like once the EU and US fast tracks catch up to their own ambitions.

What this means for sponsors

Treat the Breakthrough Devices designation and MDUFA VI capacity investment as invitations, not guarantees. A faster queue with an undefined bar is still a queue you can fail. Sponsors pursuing either pathway should build their dossiers now around the lifecycle governance elements FDA has already signaled it values for imaging AI: change control protocols, drift monitoring, and provenance documentation, then map those same elements against MDR and IVDR conformity requirements before assuming a notified body will accept less because the label says breakthrough.

Speed is being built into the system. Standards are still being negotiated inside it. The sponsors who win the fast lane will be the ones who already know what belongs in the file.

---

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 is coherent and defensible—that new fast-track pathways address throughput without specifying AI-specific evidentiary standards—but the claim that FDA imaging guidance was issued ‘despit
Source & Claim Verification	Qwen · local	cleared. All factual claims are supported by citations, but the article could benefit from more specific references to the content of the cited sources, particularly for the claims about the lack of AI-specifi
Regulatory & Framework Fidelity	Mistral	cleared. The briefing accurately reflects the procedural focus of MDCG 2025-9 and MDUFA VI but understates the implicit alignment of FDA’s imaging guidance with ISO 42001 and EU AI Act requirements for risk ma
Technical Accuracy	Llama	cleared. The article accurately describes the regulatory developments and challenges surrounding AI device review, but could be improved with more technical depth on AI-specific evidentiary standards.
Bias, Balance & Hype Control	Gemini	cleared. The briefing effectively identifies and counters potential vendor hype by highlighting the distinction between accelerated review processes and the lack of clear, AI-specific evidentiary standards.
Novelty & Non-Duplication	Grok	held. The core items (MDCG 2025-9 Breakthrough path and MDUFA VI AI capacity language) are already on the wire via the cited Jones Day and BPC pieces, and the ‘fast lanes without an evidence standard’ thesi
Validation	DeepSeek	cleared. The briefing’s central claim—that new fast-track pathways for AI devices lack new, clear evidence standards—is validated by the cited regulatory documents which detail procedural mechanics and capacit

Sources cited: 8. Validation challenges: 0. Review cost: about $0.04. Learn how these briefings are written and verified.