EU AI Act has limited impact on healthcare AI market access

The EU Artificial Intelligence Act is touted as the world’s most ambitious attempt to regulate AI, prompting fears that stricter rules could unintentionally suppress technological progress. In healthcare, where AI systems are increasingly embedded in diagnostics, monitoring, and clinical decision support, the risk of regulatory friction has been a central concern for policymakers and industry alike.

A new study published in AI & Society assesses this concern. Titled Diagnosing the Impact of the EU AI Act on Market Access and Pace of Innovation of AI-Enabled Healthcare Devices in the EU, the research evaluates whether the AI Act’s announcement coincided with measurable changes in device approvals, innovation output, or investment behavior across the EU healthcare sector.

Market access remains stable despite regulatory anxiety

AI-enabled healthcare devices already fall under some of the most stringent regulatory regimes in the world, including the EU Medical Device Regulation and In Vitro Diagnostic Regulation. Critics argued that layering a horizontal AI law on top of these frameworks would create regulatory overload, slowing approvals and delaying patient access to innovation.

To test this claim, the authors examine regulatory approval data for AI-enabled medical devices before and after the AI Act’s announcement. Using approvals from both EU-specific sources and international benchmarks, the study tracks whether manufacturers changed their behavior once the new regulatory direction became clear.

The results show no sustained reduction in the number of AI-enabled healthcare devices reaching the market after the AI Act’s announcement. While there was an observable increase in approvals following a lag period, this rise did not represent a lasting acceleration or deceleration compared to pre-announcement trends. Instead, approval activity returned to levels broadly consistent with earlier patterns.

The findings suggest that manufacturers did not immediately withdraw from the EU market or delay approvals en masse in response to the AI Act. This runs counter to warnings that companies, particularly small and medium-sized enterprises, would shift development and commercialization efforts to less regulated regions.

The study also highlights the importance of regulatory lag. Approval timelines for medical devices typically span 18 to 24 months in the EU, meaning short-term reactions to policy announcements may not be immediately visible. By accounting for this lag, the analysis avoids attributing normal regulatory delays to regulatory fear, offering a more measured view of market access dynamics.

Innovation slows, but not because of regulation alone

The study analyzes patent filings related to AI-enabled healthcare technologies to determine whether the AI Act announcement altered the pace of innovation.

Before the AI Act was announced, AI healthcare patent filings in the EU were increasing steadily. After the announcement, the growth rate slowed and eventually reversed direction, resulting in a decline in filings over time. However, the authors caution against interpreting this shift as direct evidence that regulation stifled innovation.

The post-announcement period still recorded a higher average number of filings than earlier years, suggesting that innovation momentum remained strong even as growth moderated. The slowdown appears to reflect a normalization after rapid expansion rather than a sudden regulatory shock. Broader factors, including maturing AI technologies, shifting patent strategies, and global adjustments to AI intellectual property rules, likely played a role.

The study also points to structural limitations in using patent data as a proxy for innovation. Patent filings are influenced by administrative delays, strategic behavior by firms, and evolving standards around AI inventorship and disclosure. As a result, short-term changes in patent trends may not accurately capture underlying research and development activity.

Importantly, the analysis finds no clear evidence that the AI Act announcement caused companies to abandon AI healthcare innovation in Europe. Instead, innovation appears to have continued, albeit at a more measured pace, consistent with a sector transitioning from rapid experimentation to more regulated deployment.

Investment trends defy predictions of capital flight

Critics of the EU AI Act argued that regulation would scare off investors. Venture capital plays a crucial role in scaling AI healthcare technologies, particularly for startups navigating long development cycles and high compliance costs. A sharp decline in investment would signal a loss of confidence in the EU as an innovation hub.

The study’s analysis of venture capital data tells a different story. Investment in EU-based AI healthcare startups increased in the years following the AI Act’s announcement. Both the number of deals and the total volume of funding rose more sharply in the EU than in other regions during the same period.

While the authors stress that these findings are descriptive rather than causal, they undermine claims of immediate capital flight. Instead, the data suggest that regulatory clarity may have reassured investors by signaling a stable and predictable legal environment. For healthcare technologies, where trust, safety, and compliance are critical, regulation can function as a confidence-building mechanism rather than a deterrent.

The authors also note that the surge in generative AI interest from 2023 onward likely contributed to rising investment across sectors, including healthcare. However, even accounting for this global trend, EU investment patterns do not support the view that the AI Act discouraged capital formation.

This divergence between investment growth and patent moderation highlights a key insight of the study. Innovation ecosystems do not respond to regulation in a single, uniform way. Funding decisions, regulatory approvals, and patent strategies may evolve independently, shaped by different incentives and constraints.

A more nuanced picture of regulation and innovation

The study challenges a binary framing of regulation versus innovation. In highly regulated sectors like healthcare, innovation is already shaped by legal requirements, ethical standards, and safety obligations. The AI Act did not introduce regulation into a previously unregulated space but added an AI-specific layer to an existing compliance landscape.

The authors argue that short-term analyses often overestimate the disruptive effects of regulation while underestimating the adaptability of firms. Healthcare companies, particularly those developing AI-enabled devices, are accustomed to navigating complex regulatory environments. The AI Act, while demanding, may not represent a fundamentally different challenge.

Additionally, the study does not dismiss regulatory concerns. It emphasizes that the long-term effects of the AI Act remain uncertain, particularly as enforcement mechanisms come into force and compliance costs become clearer. Small firms with limited regulatory resources may still face disproportionate burdens, and future research will be needed to assess whether innovation becomes more concentrated among large players.

Implications for policymakers and industry

The study offers evidence that early fears of regulatory collapse were overstated. The absence of a sharp decline in market access or investment suggests that the EU AI Act did not immediately undermine the AI healthcare sector. This may strengthen arguments that regulation and innovation can coexist when rules are clear and aligned with sector-specific needs.

For industry, the findings call for strategic adaptation rather than regulatory avoidance. Firms that invest early in compliance capabilities may be better positioned to compete in a market where trust and safety are increasingly central to adoption.

The study also highlights the need for better data infrastructure. Limitations in public databases, delays in reporting, and fragmented approval records complicate efforts to evaluate regulatory impact. Improving transparency around AI device approvals and innovation metrics would support more informed policy debates.

The future of AI-enabled healthcare in Europe will depend not only on the letter of the law but on how regulators, manufacturers, and investors navigate its implementation. For now, the data suggest that Europe’s AI healthcare sector is bending under regulation, not breaking.