Derbyshire Police Officer Under Criminal Investigation for Allegedly Using AI to Fabricate Evidence

A Derbyshire Police officer is under criminal investigation for allegedly using artificial intelligence to create evidential material in multiple criminal cases, Sky News reported on 13 June 2026. The Crown Prosecution Service is working alongside Derbyshire Constabulary on the inquiry. No arrests have been made.

The officer has been removed from frontline duties pending the outcome of the investigation. The allegation, as reported by The Telegraph, is that the individual used technology to manufacture evidence — conduct that, if proved, would constitute perverting the course of justice under English law. That charge carries a maximum sentence of life imprisonment and reflects the severity with which courts treat any interference with the integrity of criminal proceedings.

The involvement of the CPS at this stage is notable. Prosecutors typically enter an investigation in an advisory capacity before charges are laid, helping police frame lines of inquiry that will be sustainable at trial. Their early engagement here suggests investigators are already examining whether specific cases — and potentially specific defendants or verdicts — were materially affected by the alleged conduct.

The precise AI tooling involved has not been disclosed publicly. That gap matters, because the evidentiary implications differ significantly depending on the modality. Generative text models can produce plausible-looking witness statements or transcripts; image and video diffusion models can synthesise scene photographs or CCTV-style footage; audio models can clone voices. Each category poses a distinct forensic challenge for investigators trying to establish what is authentic in the affected case files.

Worth flagging: the detection of AI-generated evidence within existing case archives is not a solved problem. Forensic watermarking standards — such as those being developed under the C2PA (Coalition for Content Provenance and Authenticity) framework — apply prospectively to content produced by compliant tools, and offer no retrospective coverage. Statistical detectors trained to identify generative artefacts carry meaningful false-positive and false-negative rates, and their reliability degrades as underlying models improve. Any review of prior cases touched by this officer will require careful, methodology-by-methodology examination rather than a single automated sweep.

The case also lands in a specific institutional context. UK policing has been under sustained scrutiny over standards of evidence handling and disclosure — the Post Office Horizon scandal, though a different domain, sharpened public and judicial attention to what happens when institutions allow systemic failures to persist unchallenged. An individual officer acting deliberately is a different category of failure from a systemic one, but the downstream effect on affected defendants is comparable: convictions resting on fabricated material.

Looking at the broader picture, this is the kind of incident that regulatory and standards bodies have warned about in the abstract for several years. The College of Policing has been building out guidance on AI use in law enforcement, and the National Police Chiefs' Council has worked to delineate acceptable from unacceptable applications. Those frameworks were designed primarily around algorithmic decision-support tools — predictive analytics, facial recognition — not the deliberate misuse of generative models by individual officers. This case tests a gap in that architecture.

The investigation's scope — described as covering multiple cases — means the consequences could extend well beyond any sanction against the officer. Defence solicitors with clients convicted in cases the officer worked will have clear grounds to seek case review through the Criminal Cases Review Commission. The CCRC's existing caseload is already substantial; a referral wave tied to a single officer's alleged conduct could add pressure to a body with limited capacity.

Derbyshire Constabulary has not publicly detailed how the alleged conduct was discovered, over what period it occurred, or how many cases are under review. Those details will matter enormously — both to any prosecution and to the scale of the remediation effort ahead.

The criminal justice system's integrity depends on the presumption that the evidence presented in court is what it purports to be. Generative AI's capacity to produce convincing synthetic material at low cost and low skill threshold changes the threat model for that presumption, not just from external bad actors but, as this case alleges, from within.