AI Misuse in Policing: Officer Probed for Generating Evidence in Multiple Cases

For the AI community, this incident is a worst-case scenario of irresponsible deployment: a law enforcement professional allegedly used generative AI to fabricate the very material meant to uphold justice. It underscores that no institutional setting is immune from misuse, and that technical capabilities have raced ahead of policy. The timing—coinciding with the launch of a centre dedicated to ethical AI policing—amplifies the need for model transparency, use-case restrictions, and real-time monitoring in all public sector AI applications.

A Derbyshire Police officer is under criminal investigation for allegedly using artificial intelligence to create evidential material in multiple cases—the first known instance of AI-fabricated evidence in the UK criminal justice system. The case, revealed on June 13, 2026, centers on suspected perverting the course of justice. While the force has not disclosed the exact nature of the material, possibilities range from fabricated witness statements to digitally altered images or audio. The officer has been removed from frontline duties, though no arrests have been made, and the Crown Prosecution Service is actively engaging with defence lawyers and courts regarding potentially affected cases.

“

A Derbyshire Police officer is under criminal investigation for allegedly using artificial intelligence to create evidential material in multiple cases—the first known instance of AI-fabricated evidence in the UK criminal justice system.

The extraordinary timing of this disclosure coincides with the launch of PoliceAI, the UK's new national centre for AI in policing. At the launch on Wednesday, June 10, interim director Alex Murray publicly urged some forces to stop using AI systems to prepare court statements and other sensitive tasks, warning they might not be reliable. The juxtaposition is stark: a high-profile initiative promoting responsible AI adoption debuts alongside a scandal that embodies its worst fears. Murray's call for "responsible" use rings hollow when an officer—presumably aware of departmental policies—allegedly crossed the line from assistance to fabrication.

The implications for the integrity of criminal justice are profound. Evidence is the currency of the courtroom; if it can be generated by a foundation model with a few prompts, the entire adversarial system faces an authentication crisis. Defence teams now must grapple with the possibility that key documents or recordings were not merely lost or mishandled but were never authentic to begin with. The burden of proof may shift, requiring prosecution to demonstrate the human origin and chain-of-custody of every digital exhibit. This case forces an urgent reconsideration of digital evidence standards, potentially demanding cryptographic signing, watermarking, and forensic AI-detection tools as standard.

For the wider law enforcement community, the scandal is a clear warning. While UK forces have begun integrating AI for tasks like report drafting and analytics, no established framework exists to audit or certify those tools for evidence production. The absence of a formal regulatory structure—outside of generic data protection laws—leaves an accountability gap. If even one officer could allegedly misuse the tools, how many other cases might be tainted by unnoticed AI insertions? The Derbyshire investigation may only be the tip of the iceberg, as AI-generated content becomes indistinguishable from human-produced material. Police forces worldwide will be watching closely; any conviction overturned as a result could trigger a review of thousands of past cases.