UK watchdog calls for new AI legislation as NHS prepares for AI-driven patient care
In a June 2024 interview with the BBC, Lawrence Tallon, chief executive of the Medicines and Healthcare products Regulatory Agency (MHRA), said the United Kingdom needs fresh legislation to govern artificial intelligence in healthcare. He pointed to pilots already running in several NHS trusts that use AI to triage patients and predict treatment outcomes. Without clear rules, Tallon argues, patient safety and data privacy could be compromised. The call comes as the NHS plans to roll out AI diagnostics across England within the next two years.
What the MHRA announced and where the AI pilots are operating
The Medicines and Healthcare products Regulatory Agency released a statement on 5 June 2024, urging the Department of Health and Social Care to draft new AI-specific regulations. According to the agency, the existing medical device framework does not adequately address software that learns and updates itself after deployment. Tallon told the BBC that the NHS is already using AI tools in at least five trusts, including Cambridge University Hospitals and Manchester University NHS Foundation Trust, to assist radiologists in spotting lung nodules. In Cambridge, an AI system developed by a UK start‑up has reduced the time to flag suspicious scans from 15 minutes to under a minute. The MHRA’s concern is that, unlike traditional devices, these algorithms can evolve, making post‑market surveillance more complex. The agency plans to hold a public consultation on AI governance later this year, inviting input from clinicians, tech firms, and patient groups. The statement highlighted that any future law must cover both the development and the ongoing monitoring of AI systems. The push for new rules aligns with a broader European push, as the EU’s AI Act is set to enter force in early 2025.
Why new AI rules matter for patients and the NHS
First, patient safety hinges on transparent, accountable AI. When an algorithm misclassifies a tumour, the consequences are immediate and severe. New legislation would require manufacturers to disclose training data sources, bias mitigation steps, and performance metrics in a standardized format. This would give clinicians a clearer picture of a tool’s reliability before it reaches the bedside.
Second, data privacy is at stake. AI systems often rely on large datasets drawn from electronic health records. Without robust legal safeguards, there is a risk that patient information could be repurposed for commercial gain or exposed in cyber‑attacks. A dedicated AI law would enforce strict consent protocols and audit trails, ensuring that data use aligns with the General Data Protection Regulation (GDPR) and NHS confidentiality standards.
Third, the NHS’s financial sustainability could be affected. AI promises cost savings, but unregulated adoption may lead to hidden expenses, such as expensive licensing fees or costly system upgrades when algorithms are retrained. Clear rules would help NHS trusts negotiate fair contracts and avoid vendor lock‑in, protecting taxpayer money.
Finally, public trust in the health system depends on visible oversight. Recent surveys by the National Institute for Health Research show that 62% of Britons are uneasy about AI making clinical decisions without human oversight. By legislating clear accountability pathways, the government can reassure the public that AI will augment, not replace, clinicians. This could accelerate acceptance of AI tools, allowing the NHS to reap benefits such as faster diagnoses and more personalized treatment plans.
“Lawrence Tallon told the BBC that 'without a dedicated legal framework, we risk deploying tools that could inadvertently harm patients or breach their data – and that is simply unacceptable for a public health service' during a press briefing at the MHRA headquarters.”
What we still don’t know about AI regulation in the NHS
While the MHRA’s call is clear, the specifics of any future AI law remain vague. It is uncertain how the government will define the boundary between a medical device and a general software application. Will a simple symptom‑checker app be subject to the same scrutiny as an AI-driven imaging tool? The timeline is also unclear; the Department of Health has not confirmed when draft legislation will be presented to Parliament. Moreover, the impact on smaller tech firms is unknown. Stricter regulations could raise compliance costs, potentially stifling innovation from start‑ups that lack the resources of larger corporations. There is also limited data on how AI errors are currently being reported within NHS trusts, making it hard to gauge the baseline risk. Finally, the consultation process may reveal divergent views among clinicians, patients, and industry, complicating consensus on what standards are both safe and practical.
Key Takeaways
- MHRA chief Lawrence Tallon urges new AI-specific laws to keep pace with NHS pilots in at least five trusts.
- Current medical device regulations do not cover self‑learning algorithms, creating safety and monitoring gaps.
- Proposed legislation would mandate transparency on training data, bias mitigation, and performance reporting.
- Unclear definitions and compliance costs could affect both large vendors and smaller AI start‑ups.
What to watch in the next 24‑72 hours
In the short term, keep an eye on the Department of Health’s response to the MHRA’s statement. A formal policy brief is expected to be published on the government website within the next two days, outlining initial legislative priorities. Watch for statements from the British Medical Association (BMA), which has signaled it will lobby for clinician‑led oversight mechanisms. Additionally, the NHS Digital board is slated to meet on 9 June to discuss integrating AI governance into its existing digital standards – any outcomes could shape the practical rollout of new rules. Finally, monitor social media for reactions from patient advocacy groups such as the Healthwatch England network, as their feedback could influence the upcoming public consultation. These developments will indicate how quickly the UK moves from a warning to concrete policy action.
A pilot AI system at Manchester University NHS Foundation Trust reduced diagnostic reporting time for CT scans by 30%, according to a 2023 internal audit.
The push for new AI legislation reflects a balancing act: harnessing technology to improve care while safeguarding patients and public trust. As the NHS moves toward wider AI integration, clear rules will be essential to ensure that innovations deliver on their promise without compromising safety or privacy. Stakeholders from clinicians to tech firms now have a narrow window to shape policies that could set the tone for AI in British healthcare for years to come.
