Judge said advocates agreed AI-generated summaries were “immensely useful” in improving accessibility
A Family Court judge has used generative artificial intelligence to produce a simplified explanation of a ruling for parents with learning and cognitive difficulties, in what is understood to be one of the first recorded instances of a judgment being explicitly supported in part by AI.
The case, Re B (Fact Finding: Use of AI for Judgment Summary for Parents with Learning Difficulties), was heard by Her Honour Judge Hesford, who appended two additional versions of her judgment with the agreement of all parties and their advocates.
The supplementary sections were produced from the full judgment using a secured judicial AI application based on Microsoft Copilot, following prompts to simplify and summarise the court’s reasoning. The material was designed to assist the parents in understanding the outcome of proceedings, given their learning and cognitive difficulties.
The judge produced both a simplified narrative summary and a bullet-point version of the judgment to improve accessibility for the parties.
In her notes, HHJ Hesford said: “All advocates agreed that they were immensely useful and they have been attached specifically to show how useful AI can be when carefully used in such situations.”
The underlying proceedings involved a five-day fact-finding hearing concerning a child, referred to as B, now aged two, following a skull fracture suffered at eight months while in the care of her biological parents. A local authority, not named, had applied for a care order, which was opposed by the married parents.
The court heard that the baby was taken to hospital after developing swelling to her head and around her eye. While neither parent disputed the medical findings regarding the injury, both denied causing it and said it was accidental.
In her findings, the judge concluded that inconsistencies in the parents’ evidence were attributable to confusion, limited understanding, and their individual difficulties rather than deliberate deception.
She did not find, on the balance of probabilities, that the injury had been deliberately inflicted. However, she concluded that it resulted from an event involving one of the parents and that the standard of care fell below what would reasonably be expected. The judge was unable to identify the perpetrator, and both parents remained within the pool of possible perpetrators.
The case will now proceed to the welfare stage, where the court will determine the child’s long-term future. The judge also noted that the proceedings had already been ongoing for more than 18 months, highlighting delay within the process.
The use of AI in the case was limited to producing explanatory summaries and did not form part of the judicial decision-making process.