Artificial intelligence-enabled triage systems, digital consultation platforms, ambient voice technology, clinical decision-support tools and remote monitoring solutions are becoming part of everyday care delivery. At the same time, healthcare organisations are facing growing scrutiny and pressure through patient safety investigations, complaints (increasingly themselves drafted using AI), regulatory processes and evolving coronial systems.
There have been several high level publications recently on the use of AI, in particular, in September 2026, the National Commission into the Regulation of AI in Healthcare report whose recommendations were backed in full by the government, followed by guidance on AI from NHS Resolution. Within these, important questions were raised around safety, accountability, governance and public confidence and AI. Although developed in the context of wider healthcare systems, many of the underlying themes are relevant to coronial investigations.
As AI becomes more embedded in clinical care, reported coronial decisions and Prevention of Future Deaths reports containing AI concerns seem likely to start appearing.
What role will AI play at inquests?
The arrival of AI does not remove the need for sound clinical judgement, effective governance or clear decision-making. However, it may create new evidential and practical challenges when those issues are scrutinised after a death. Questions may arise such as:
what role did the AI system play in the patient's care?
was the output advisory or determinative?
what level of human oversight was expected?
were staff appropriately trained in the use of the technology?
what safeguards existed to identify inaccurate or inappropriate outputs?
were concerns about performance, reliability or safety identified before the incident?
These questions could be directed at frontline clinicians but also at governance leads, digital teams, patient safety officers and those responsible for implementation and oversight.
Traceability may become as important as the technology itself
A challenge for healthcare organisations is likely to be demonstrating precisely what role an AI system played in a patient's clinical journey.
Healthcare providers routinely retain clinical records, policies and investigation materials. However, AI-assisted systems may require organisations to go further as it may not be sufficient to simply establish that a particular technology was used. Providers may also need to identify which product was deployed, which version was operating at the time and what influence, if any, it had on subsequent decision-making.
The National Commission report recognises that AI systems may evolve over time and, as such, supports:
lifecycle regulation rather than relying solely on point-in-time approval
continuous monitoring of AI-enabled technologies throughout their operational life
recognition that AI systems may change over time
greater emphasis on post-market surveillance and real-world performance monitoring
Providers should therefore consider whether their governance arrangements and audit trails would enable them to reconstruct the role played by an AI tool months or even years after an event. The organisations best able to answer those questions are likely to be in a much stronger position when responding to investigations and inquests.
Ambient voice technology and the clinical record
One area likely to attract particular scrutiny is ambient voice technology (AVT) and AI-assisted documentation.
These tools can perform a range of functions, including transcribing consultations, generating summaries, structuring information within clinical records, suggesting clinical codes and preparing draft correspondence for professional review.
The potential benefits are significant. However, they also raise important evidential questions.
If an inquest concerns an alleged failure to recognise deterioration, communicate concerns or document clinical assessments appropriately, attention may extend beyond the finished record itself. Questions may arise about how information was generated, what human review occurred, whether changes were made before clinician approval and whether relevant information was omitted, misunderstood or inaccurately recorded.
Healthcare providers should be able to demonstrate not only that appropriate policies existed but also that clinicians understood their continuing responsibility for reviewing, editing and approving AI-generated outputs before they became part of the clinical record.
Disclosure may become more complicated
Healthcare organisations are already familiar with extensive disclosure obligations in serious incident investigations and inquests, but AI may broaden the range of potentially relevant materials. Depending on the circumstances, healthcare providers may now also need to locate and preserve:
AI-generated records, summaries or recommendations
audit trails and system logs
implementation records
governance papers
risk assessments and hazard logs
clinical safety documentation
staff training materials (on the purpose of the tool, the data it uses and how to use the tool in conjunction with their processes and what communications are necessary with the patient)
monitoring and assurance documentation; and
data protection or information governance assessments
The ability to demonstrate not only what the system did, but what the organisation understood about its risks and how those risks were managed, may become increasingly important.
Who is responsible?
An inquest does not determine civil liability. Nevertheless, questions of responsibility will inevitably arise where clinical decisions have been influenced by AI-generated outputs.
The guidance recently issued by NHS Resolution reminds NHS organisations that existing duties of care and negligence principles continue to apply where AI-enabled technologies are used in patient care. While manufacturers may have responsibilities where harm arises from a defective product, healthcare providers and clinicians remain responsible for the delivery of safe care and for exercising appropriate clinical judgement.
Recent commentary from clinical defence organisations has highlighted concerns that clinicians may become the obvious focus of scrutiny when AI-assisted decision-making contributes to an adverse outcome, even where they have limited visibility of how a system reached its conclusions.
Coroners are unlikely to be interested in abstract debates about liability allocation. They are, however, likely to explore how clinical judgement, governance arrangements, organisational decision-making and technological systems operated together in practice.
Looking ahead
The organisations best placed to respond to future scrutiny will not necessarily be those with the most advanced technology. They will be those able to demonstrate robust governance, clear accountability, effective audit trails, meaningful oversight and a clear understanding of how AI influenced a patient's care.
For further assitance on AI and the coroner's court, please contact our expert health and care solicitors.
Read More
