resultsense.com web signal

MPS warns NHS doctors left liable for AI errors

healthcare microsoft nhs ai-liability healthcare medical-ai

Key insights

  • The MPS is calling for AI tools to be reclassified as products under the Consumer Protection Act 1987, shifting liability away from clinicians to developers.
  • Under current UK law, doctors can be sued for AI errors they did not cause, including missed tumours on X-rays and wrong warfarin dosing.
  • Microsoft Copilot is deploying to more than 500,000 NHS staff while NHS Resolution is still drafting AI liability guidance.

Why this matters

The NHS is already deploying AI at scale with more than 500,000 staff on Microsoft Copilot, but no settled legal framework yet determines who is liable when clinical AI fails. For AI developers building medical tools in the UK market, a reclassification under the Consumer Protection Act 1987 would directly shift their risk exposure, insurance requirements, and product design obligations. NHS Resolution drafting liability guidance in parallel with mass deployment means the accountability rules are being written while the technology is already in clinical use.

Summary

The Medical Protection Society has warned UK ministers that doctors are the "liability sink" for AI-caused medical errors under current law. The MPS cites two scenarios: AI missing a tumour on a chest X-ray, or recommending the wrong warfarin dose triggering severe bleeding. In both, clinicians could be "held wholly liable" for clinical negligence. Its fix: reclassify AI tools as products under the Consumer Protection Act 1987, shifting liability to developers. Essentially: (MPS, NHS Resolution) are behind the curve, with Microsoft Copilot already deploying to more than 500,000 NHS staff. - NHS Resolution is still drafting AI liability guidance, the Department of Health confirmed. - The Health Foundation warns clinical AI trust could erode faster than rollout. The tools already read scans, summarise consultations and draft patient letters. The legal framework for when they fail is not yet written.

Potential risks and opportunities

Risks

  • NHS clinicians using AI diagnostic tools could face personal negligence claims for AI-caused missed diagnoses before NHS Resolution's liability guidance is finalized
  • Microsoft faces reputational and regulatory exposure if Copilot errors surface in clinical settings while accountability frameworks are still being drafted
  • The Health Foundation warns public trust in clinical AI could erode faster than the tools are rolled out if liability questions remain unresolved

Opportunities

  • Legal and insurance firms with technology product liability expertise could see growing demand as NHS Resolution finalizes guidance covering more than 500,000 NHS AI users
  • AI developers who voluntarily adopt product liability standards ahead of any Consumer Protection Act 1987 reclassification could reduce friction in NHS procurement as governance scrutiny rises
  • The MPS and Society for Acute Medicine's aligned pressure on governance creates a policy window for health AI vendors to help shape UK liability frameworks before they are codified

What we don't know yet

  • Whether NHS Resolution's draft liability guidance will align with the MPS's proposed Consumer Protection Act 1987 reclassification or adopt a different legal framework
  • Timeline for when the Department of Health expects NHS Resolution to publish finalized AI liability guidance for NHS clinicians
  • Whether the more than 500,000 NHS staff now on Microsoft Copilot have any interim liability protections while formal guidance remains pending