Innovation vs. Regulation

Innovations, yes. What about regulations?

Artificial intelligence (AI) is revolutionising healthcare, promising unprecedented improvements in diagnostics, patient care and administrative efficiency. However, the UK currently faces significant challenges in regulating these fast-evolving technologies, with governance frameworks lagging behind innovation.

A fragmented regulatory landscape

Unlike the European Union's comprehensive AI Act, the UK's regulatory approach is fragmented across multiple sector-specific agencies. The Medicines and Healthcare products Regulatory Agency (MHRA) oversees AI-based medical devices, focusing on principles of safety, security, robustness, transparency and fairness. In parallel, the Information Commissioner’s Office handles data privacy, creating a complex landscape for healthcare providers navigating conflicting guidance.

The AI Airlock pilot by MHRA represents a notable attempt to streamline regulatory practices, testing real-world applications of AI medical devices. However, isolated initiatives struggle to tackle systemic challenges such as algorithmic bias, which transcends individual regulatory domains.

The challenge of adaptive AI systems

Current UK regulations were designed for static medical technologies. Adaptive AI systems that evolve through continuous learning, like those predicting conditions such as sepsis, require costly and time-consuming re-validation after each update under current regulations. This poses a significant barrier to innovation, forcing developers to limit adaptive capabilities or face regulatory non-compliance.

Accountability frameworks also remain inadequate, especially regarding algorithmic drift, creating uncertainty around liability when AI systems recommended by NICE guidelines inadvertently cause patient harm.

Data governance and health inequalities

While GDPR sets baseline protections, it falls short in addressing the nuanced needs of AI training processes, particularly regarding synthetic data use. The NHS's fragmented data infrastructure exacerbates these issues, resulting in AI systems that perform inconsistently across geographic regions, potentially reinforcing healthcare inequalities.

Innovation vs safety: striking the balance

The UK government's pro-innovation stance has enabled rapid AI deployment, as seen during COVID-19 with AI-driven triage systems. However, this has occasionally compromised patient safety, highlighting the necessity for balanced regulation. The ongoing debate around the proposed AI Regulation Bill 2025 underscores this tension, with proponents advocating central oversight to mitigate risks like the Babylon Health chatbot incidents, and critics concerned about stifling innovation and economic growth.

Workforce preparedness and implementation gaps

Regulations currently overlook essential human-AI interaction dynamics. A significant number of healthcare professionals report inadequate training in interpreting AI diagnostics, leading to inconsistent adoption and potential patient safety risks. The UK also lacks explicit plans to address potential workforce displacement by administrative AI tools, unlike the EU’s proactive employment impact assessments.

International alignment challenges

Post-Brexit divergence from EU regulations has introduced additional compliance burdens, particularly for healthtech firms operating across UK-EU borders. This regulatory fragmentation threatens critical infrastructure investments and risks isolating the UK from emerging international regulatory standards.

Emerging solutions and future directions

Promising initiatives, such as MHRA’s AI Airlock and NICE’s forthcoming AI-specific evaluation frameworks, demonstrate potential pathways forward. However, significant challenges persist:

• Balancing centralised governance with sectoral expertise

• Creating regulations flexible enough to accommodate evolving AI technologies

• Reconciling innovation leadership with rigorous patient safety standards

The next 12-18 months are critical. With the EU AI Act effective from 2026 and the US FDA updating its digital health framework, the UK must urgently align its regulatory standards to remain globally competitive. Successfully bridging this regulatory gap is essential to harness AI’s full potential safely and effectively within UK healthcare.

Connect with Contempo and join the discussion!