The EU’s Dual Regulatory Challenge for AI Diagnostics
AI diagnostics are the next regulatory frontier for medical technology. Existing medical device rules under MDR and IVDR already created a rocky road for traditional products. The EU AI Act now adds a separate layer by classifying most clinical AI as high-risk, creating overlapping obligations for the same product.
The Burden on AI Innovation
Manufacturers face a double layer of requirements: device conformity under MDR/IVDR plus AI Act mandates on risk management, transparency, documentation and post-market monitoring. That duplicates audits, multiplies technical documentation and raises costs. Small and medium-sized enterprises, university spin-outs and early-stage investors are most exposed. The likely result is slower market entry, fewer choices for clinicians and delayed patient access to valuable diagnostics.
Pathways to Streamlined AI Regulation
Several practical measures can reduce unnecessary friction while protecting patients. First, adopt a least burdensome approach: calibrate documentation and testing to the novelty and clinical risk of the algorithm rather than applying full device-level burdens to minor model updates. Second, expand regulatory sandboxes that let developers test algorithms in controlled clinical settings with regulator input and agreed evidence requirements. Third, apply Well-Established Technology (WET) principles so validated components and mature models face proportionate scrutiny.
Operational alignment is also essential: harmonise conformity assessment processes so MDR/IVDR audits and AI Act checks share evidence and avoid duplication. Finally, modernise the regulatory workforce by hiring data scientists, ML engineers and algorithm auditors capable of assessing model drift, training data provenance and bias mitigation.
Securing Future AI Health Solutions
The objective is not deregulation but calibration. A coherent, coordinated EU approach that links MDR/IVDR procedures with AI Act rules, promotes sandboxes and proportional assessment will protect patients while keeping Europe competitive for AI diagnostics. Without timely alignment, Europe risks safer technology remaining unavailable to clinicians and patients who need it.
