Derbyshire Police Officer Investigated Over AI-Created Evidence
TL;DR
- A Derbyshire officer faces criminal investigation for allegedly using AI to draft victim impact statements and prosecutor briefings in rape cases.
- Multiple rape convictions are reportedly under review, though the CPS has not disclosed how many cases are affected.
- PoliceAI's head asked some forces to pause AI use in court document preparation, days after the national centre launched on 10 June 2026.
When a police officer uses AI to generate the victim impact statements and prosecutor briefings that drive charging decisions in rape cases, the problem is not the AI going wrong on its own; it is a human deliberately steering it toward a preferred outcome. The Guardian reported that a Derbyshire officer is under criminal investigation on suspicion of perverting the course of justice after allegedly doing exactly that. The officer, removed from frontline duties, reportedly prompted the software to maximise the apparent impact of statements in order to secure charges, according to multiple reports. No arrests have been made.
Multiple rape convictions are reportedly under review as a result. The Crown Prosecution Service has confirmed it is engaging with defence lawyers and the courts over affected cases, though the number of convictions under scrutiny has not been disclosed. That silence on scale is, for now, the most important open question.
This is believed to be the first criminal investigation of its kind in the UK, and the timing is notable. PoliceAI, a new national centre backed by £75 million in Home Office funding and intended to coordinate AI deployment across 43 forces in England and Wales, formally launched just days before the story broke. Alex Murray, who heads PoliceAI, has since asked some forces to pause AI use in preparing court documents while safeguards are developed, and has stated that technologies used within the criminal justice system must achieve a standard of accuracy that is "beyond reasonable doubt." That framing is pointed: generative AI systems are prone to fabricating details, a critical flaw when such material enters legal proceedings.
What the reporting does not yet give you is the full picture. Which AI tool the officer used, whether that use was authorised by force policy, and whether supervisors or prosecutors reviewed and approved these documents without detecting their AI origin are all still unknown. If internal review chains failed to catch the problem, individual prosecution is unlikely to fix that process failure on its own.
The PoliceAI centre now has a real stress test before it has properly launched. Murray's early intervention is exactly the kind of course correction a national oversight body should provide. Whether it arrives in time to set binding standards before this investigation discloses its full scale is the question the coming weeks will answer.
Shared on Bluesky by 3 AI experts
-
www.theguardian.com/technology/2...
View on Bluesky →
Originally reported by theguardian.com
Read the original article →Original headline: Derbyshire police officer investigated over AI-generated ‘evidential material’