AI in inquests – risks for healthcare professionals and their indemnifiers

The use of Artificial Intelligence (‘AI’) in all kinds of litigation is increasing. In a September 2025 survey, over 60% of UK lawyers said they were using generative AI for legal work. No reliable statistics exist about how many litigants in person are using AI, but anecdotal evidence suggests that it is already being widely used.

In this Keynote, Tracy Sell-Peters and Joanne Staphnill consider the risks for healthcare professionals and their indemnifiers when AI is used in the coronial process, and how these risks can be managed.

AI used where expert not permitted

An example has been reported of AI being used in an inquest when the coroner did not permit an expert. In that case, a man in his 70s died unexpectedly after complex cardiac surgery. A barrister was instructed by his family in the inquest. The coroner’s duty includes establishing how the deceased came by their death, and the family’s barrister recognised that this would involve understanding potentially complex medical issues.

The barrister’s request for an independent expert report was denied by the coroner for reasons which have not been reported. Expert input is usually sought so the coroner has assistance with complex medical, technical, or scientific issues, especially those relevant to the cause of death. But there are reasons why a coroner might not permit expert evidence. Depending on what other evidence is available, the coroner could deem it unnecessary. The cost of using an expert is also an important consideration for coroners. If the coroner instructs an expert, or adopts an expert report originally commissioned by someone else, the (rather overstretched) coroner’s court has to pay that expert’s fees. Therefore, it is not unusual for an inquest hearing to proceed without expert evidence, even where the death follows complex medical care.

The barrister was interviewed about the case, including about his preparation to cross-examine the cardiac surgeon, without the benefit of an expert report to identify the issues. The barrister was reported as saying, “My use of ChatGPT allowed my questions to be more focused on the technical aspects of the surgery and help fill the gaps left by not having experts to call upon.” The barrister rightly emphasised that he had not put any personal information or data into the AI tool, and had also scrutinised the information generated.

Risk of AI causing harm in an inquest

The barrister in this case study was clearly taking proper care to discharge his professional duties. There is no suggestion that his use of AI to support his legal work caused any difficulty or injustice in that case.

However, the use of AI could cause significant problems in inquests, including the risk of delay or even unjust outcomes.

The usefulness and accuracy of any AI output depends on the quality of the input – the information, prompts, and questions provided by the user of the AI tool – and the quality of the source material. Also, AI tools are known to occasionally hallucinate outputs and refer to things that do not exist. Therefore, if anyone involved in the inquest is using AI, there is a risk that questions or submissions could be misconceived, unnecessary, or misleading. Information or evidence could be submitted that appears plausible but is actually not relevant, lacking in proper context, or unintentionally false. The risk may be particularly acute where some of the Interested Persons (an “IP” – the family of the deceased will usually be an IP) are unrepresented or represented by lawyers without the necessary expertise to scrutinise any AI outputs properly.

The most likely risk is wasted time and delay. Coroners are not typically medically qualified. Most are lawyers, and all receive thorough training including on how to deal with evidence. If an IP is medically qualified, and particularly if they are properly represented, they would see the inquest bundle and should spot when another party has (unknowingly) submitted misleading or false content generated by AI. Scrutiny from the coroner and medically qualified IPs should minimise the risk of inappropriate AI-generated material influencing the coroner’s findings. But by the time AI-generated errors are identified, time has already been wasted in controlling the evidence or discussing the permissibility of questions, or in reading and considering material that ultimately has to be disregarded. Coroners’ courts are already under significant pressure, and there are backlogs of inquests waiting to be heard. The time available to prepare for an inquest and for the inquest hearing itself is already limited. While AI could expedite litigation and clear backlogs, in the short to medium term significant time could be wasted through the coroner and the parties dealing with disputes over AI-generated questions, submissions, information, or purported evidence.

A risk also remains that AI-generated errors are not identified and corrected, especially if there is no expert witness available. If the death followed medical care, then some of the factual witnesses would be medically qualified. But witnesses are not always called to a hearing, and do not see the whole inquest bundle or submissions. Witnesses cannot make submissions themselves. Therefore, even medically qualified witnesses might not be in a position to correct or even see any misleading AI-generated material. If false content goes unnoticed and is accepted without proper scrutiny, it could lead to inaccurate findings by the coroner. It would be a tragedy for any family to go through the stress of an inquest and end up with inaccurate answers or findings and conclusions. These could also hamper the ability to learn from adverse events and prevent future deaths.

The risk of inaccurate findings is also concerning for individual healthcare professionals involved as witnesses or IPs. Inaccurate inquest findings could encourage the family of the deceased to start compensation claims that would not otherwise have been made. Some healthcare professionals, like doctors, are obliged to self-refer to their regulator if criticised in an inquest, so inaccurate findings in inquests could result in unnecessary regulatory investigations.

Managing the risks

It is important to get legal advice from the outset. Coroners ask clinicians for statements at the start of the inquest process, often in order to decide important questions like whether to hold an inquest, who to call as witnesses, and who needs to be made an IP. This is sometimes done before key information such as a post-mortem report is available. The difference between the positions of witnesses and IPs means that the question of whether a healthcare professional should seek IP status is always a crucial strategic question, often requiring legal advice. Legal advice is often needed to finalise that statement in a way that provides the information the coroner needs but also protects the healthcare professional or provider from misinterpretation or unwarranted criticism.

Expertise is also needed at Pre-Inquest Review hearings to make appropriate submissions about when expert evidence is essential and maximise the prospects that the coroner and parties have everything they need to make accurate findings, without unnecessary difficulty or delay.

At the hearing, it is often impossible to know in advance what questions will be raised in cross-examination, and it may only become apparent during the hearing itself if a party is basing their approach on misguided AI outputs. Therefore, specialist legal advice is essential to adapt to difficult situations like this.

If you are a medical professional, clinic or hospital and would like advice or representation in relation to an inquest into a patient’s death, please contact Tracy Sell-Peters or Joanne Staphnill.

Tracy and Joanne also advise insurers or indemnifiers who arrange and fund representation for medical professionals or providers in inquests.