UK police watchdog warns on AI use in investigations

In short: The UK police watchdog says police in England and Wales need clear rules for using AI in investigations, or cases could collapse.

What happened

Rachel Watson, director-general of the Independent Office for Police Conduct (IOPC), warned that misuse of AI could lead to mistrials and even criminals avoiding conviction. A mistrial is when a trial is stopped or ruled unfair, so the result cannot stand.

Watson said AI can sometimes “make up” information, which is often called a hallucination (like a tool confidently giving a wrong answer). If that incorrect information slips into evidence or court paperwork, it can undermine a case.

Her comments follow a case in Derbyshire Police, where a detective constable is under criminal investigation for allegedly using AI to “create evidential material”. People familiar with the case said the officer asked a chatbot to draft documents designed to increase the chance of conviction and a severe sentence. The officer has been suspended from frontline duties.

The IOPC said it has received four referrals linked to AI misuse and has two live investigations involving AI issues. One involves a Derbyshire detective chief inspector who allegedly used AI to keep legally important decision logs. The other examines how West Midlands Police used an evidence dossier that included false facts produced by Microsoft Copilot when recommending restrictions on visiting football fans.

Why it matters

Police forces are under pressure and are looking at AI to reduce paperwork and speed up investigations. But without clear guidance and oversight, AI could be used in ways that damage public trust, risk unfair outcomes in court, and expose officers to misconduct action for using tools they were never authorised to use.

Source: Financial Times