BRIDGING INNOVATION AND REGULATION
Designing Living Labs for Safe AI and Digital Health Solutions
This interview explores how Living Labs can bridge innovation and compliance, balancing flexibility with safety guardrails. By integrating smart planning and iterative adaptation, the authors propose a framework for digital health development, including AI-enabled medical devices that incorporates agile real-world testing environments while remaining within regulatory boundaries.
Theme 1. Contribution of Living Labs in Validation of AI-enabled Devices
How do Living Labs uniquely support the real world development and validation of AI enabled medical devices compared to conventional clinical trials and how can they ensure effective validation across diverse populations despite being localized or context specific?
Living Labs provide a flexible, semi-controlled environment where AI-enabled digital health technologies (DHTs) can be trialled in clinical settings that resemble everyday healthcare practice. Unlike traditional clinical trials—which are often strictly controlled, narrowly scoped, and limited to predefined protocols—Living Labs can enable iterative testing by allowing real-time feedback from users, including patients and healthcare professionals.
This setup is particularly suited to identifying challenges in device usability, clinical workflow integration, and patient interaction that would otherwise go undetected in conventional studies. Because devices are evaluated in their intended contexts of use, developers gain early insight into real-world performance and limitations.
The Living Lab model can support broader validation through the inclusion of diverse user groups and the possibility of networking across multiple facilities. Additionally, Living Labs could be used to assess how a product already approved in one jurisdiction performs in another, thus providing comparative post-approval evidence without launching a full new trial.
Theme 2. Navigating Regulatory Compliance with Flexibility
How can Living Labs enable iterative, user-driven AI development for medical devices while ensuring compliance with EU MDR/IVDR and AI Act requirements, addressing adaptive AI’s regulatory challenges, managing version control, documentation and traceability, and proactively mitigating ethical issues like bias and data privacy?
The article highlights that effective integration of Living Labs within regulatory frameworks (including in the EU) requires a structured yet adaptable approach. A central challenge in testing adaptive systems lies in maintaining traceability and ensuring that each system iteration is adequately documented and evaluated. Strategies are described to embed compliance within the Living Lab infrastructure by using digital quality management systems (QMS) that automate documentation, version control, and traceability. These systems could serve maintaining a record of changes while supporting agile development.
One of the key strategies is to design studies around a broader evaluation concept, allowing for predefined, bounded changes to the device during testing. This method—comparable to the concept of Predetermined Change Control Plans (PCCPs)—defines the scale and type of permissible adaptations without necessitating full protocol amendments. When these boundaries are well justified and transparent, ethics committees and regulatory authorities may find it more suitable to approve such flexible approaches.
Evidence generated in Living Labs should be captured through high-integrity data management tools, and ideally, mapped to regulatory expectations from the outset. This helps ensure that the insights gathered are not just valuable for development but can also be used in clinical evaluation submissions. The article stresses that applicable regulatory and ethical frameworks still apply including MDR Articles 62 (clinical investigations) and 82 (non-interventional device testing), the GDPR and those of Good Clinical Practice. It also takes into account the new requirements of the EU AI Act, including recent provisions around transparency and AI literacy.
Furthermore, it is critical that any form of testing does not infringe of the fundamental rights of individuals and prioritises informed consent and user safety. Ethical safeguards in Living Labs extend beyond traditional protocols. The roadmap recommends implementing dynamic and hybrid consent mechanisms that give participants greater control over what data are collected, when, and how they are used. This is particularly important in settings that use contactless or passive monitoring technologies, where transparency is critical to maintaining trust.
Theme 3. Managing Risk and Ensuring Safety in Semi-Controlled Environments
How can Living Labs support agile AI-driven medical device development by adapting the device's intended purpose during iterative cycles, ensuring compliance with EU MDR/IVDR, integrating post-market clinical follow-up (PMCF) for long-term safety monitoring, and implementing effective risk mitigation strategies when involving patients in early-stage testing?
Living Labs are primarily intended for use in the development and pre-market validation phases of medical devices, and depending on the technology readiness level of the device or concept being tested, the intended purpose largely depends on the scope of the clinical study defined. There can also be situations where Living Labs can be employed for post-market clinical follow-up (PMCF) where further real-world evidence is required on an already approved product.
When it comes to involving human participants—including in early-stage AI testing—Living Labs are not inherently riskier than other clinical research settings. In fact, they can enhance safety oversight by enabling close, context-aware observation. However, any interaction with patients must follow the same ethical and legal safeguards required for clinical investigations, including ethics committee approval, risk-benefit assessments, and transparent informed such as those outlined in MDR – eg Article 62 (clinical investigations).
Risk can be managed through careful planning, adherence to good clinical practice, and use of digital monitoring technologies that can detect adverse events in real time. Continuous safety monitoring and a clear escalation protocol are essential, especially when testing technologies that are not yet fully validated.
Theme 4. Enhancing Regulatory Collaboration and Scaling the Living Lab Model
How can early involvement of Notified Bodies within the Living Lab lifecycle, coupled with a collaborative roadmap among industry, academia, and regulators, enhance regulatory outcomes and help establish the Living Lab model as a standard framework for aligning regulatory science with safe, agile innovation in AI and digital health across Europe?
The roadmap envisions a future where Living Labs become a foundational element of regulatory science, bridging the gap between agile innovation and stringent safety requirements. However, this will require a commitment from stakeholders across sectors. Investment in infrastructure, development of shared standards, and expansion of regulatory capacity are all critical to the scalability of this model. Engaging regulatory authorities in the development process can facilitate iterative development, regulatory learning and helps align study designs with approval standards from the outset.
While Living Labs are not yet a standard or required element of EU regulatory frameworks, their growing adoption suggests they may become standardized in the future. As this occurs, it is essential that their core attribute—operational flexibility—is preserved.
To fully realize the potential of Living Labs, the EU must match its regulatory ambition (e.g., the AI Act) with equal investment in enabling infrastructure. This includes support for experimentation spaces that can handle the complex, real-time demands of modern AI systems without compromising safety or trust. A balanced, well-funded strategy will allow digital health technologies to be both innovative and regulation-ready, ultimately benefiting patients, providers, and healthcare systems alike.
Theme 5: Technology Readiness Levels
How are technology readiness levels (TRLs) aligned with regulatory milestones in your proposed Living Lab framework?
Technology Readiness Levels (TRLs) are used in the roadmap to determine the appropriate level of regulatory and ethical oversight. Early TRL stages (e.g., TRL 3–4) allow for exploratory, low-risk research with more flexible designs, while mid-to-late stages (TRL 6–9) require increasingly rigid controls and documentation to support regulatory approval. At each stages, appropriate ethical and regulatory interactions are described according to the intended study. This scaling of oversight allows the Living Lab to function as a versatile environment, adaptable to different stages of product maturity.
