A Metropolitan Police officer reportedly used an artificial intelligence chatbot to help assess a bone discovered during the continuing search for the remains of Muriel McKay, raising questions about the appropriate use of general-purpose AI in forensic investigations.
The development was reported by The Times on November 12, 2026, citing court documents connected to a legal challenge brought by McKay's family.
McKay was abducted in London in 1969 and subsequently murdered. Although two brothers, Arthur and Nizamodeen Hosein, were convicted of her murder, her remains have never been recovered.
The latest dispute concerns bone material discovered during searches associated with the historical investigation.
According to the report, contractors found a bone in a garden in Bethnal Green, London, in March 2026. Additional smaller fragments were recovered the following month.
Court documents reportedly indicate that a police officer uploaded a photograph of the bone to Microsoft Copilot, a general-purpose AI assistant.
The system suggested that the object was a cattle bone.
However, identifying skeletal material from photographs presents significant scientific limitations.
Forensic anthropologists normally examine anatomical morphology, measurements, preservation characteristics, and other diagnostic features to determine whether remains are human or non-human.
Photographs may support preliminary assessment, but their usefulness depends on image quality, scale, viewing angle, and the anatomical structures visible.
General-purpose AI systems are not substitutes for validated forensic anthropological examination.
The Metropolitan Police maintain that an external forensic assessment concluded the recovered material was non-human.
The available reporting does not establish that the chatbot's response formed the basis of that professional conclusion.
McKay's family is seeking a judicial review of the decision not to undertake further testing or excavation.
The family argues that the recovered fragments require additional specialist consideration and that the search location remains significant to the investigation.
Police have indicated that they are defending the legal challenge while remaining open to reviewing new evidence.
The dispute illustrates an increasingly important distinction in forensic science: AI-generated suggestions and scientifically supported expert conclusions are not equivalent.
A chatbot may produce a plausible description of an object without demonstrating the validated accuracy, reproducibility, or evidential reliability required in forensic casework.
At the same time, the reported use of AI does not independently establish that the subsequent forensic examination was incorrect.
Determining whether additional testing is justified requires consideration of the physical evidence, examination records, investigative context, and available specialist expertise.
The case highlights the importance of transparency when AI tools are introduced into forensic workflows, particularly when their outputs could influence investigative decisions.