The Independent Office for Police Conduct (IOPC) in Britain has opened a formal investigation into a senior detective in the Derbyshire Constabulary, following allegations of gross misconduct resulting from his reliance on the artificial intelligence tool "Microsoft Copilot" to assist in making and justifying decisions related to ongoing criminal investigations.
Details of allegations and violation of operational controls
The investigation began in August following a voluntary referral initiated by the Derbyshire Constabulary leadership on June 23, after an internal review revealed a pattern of the officer's use of artificial intelligence. Data indicates that the violations attributed to the detective went beyond routine tasks such as linguistic review or spell checking; he used the tool to draft procedural decisions, assess and manage risks, and justify critical decisions during criminal investigations.
The list of charges also includes entering sensitive operational data and personal information related to parties in cases into the tool's cloud system. Derbyshire Constabulary had officially permitted its members to use "Copilot" in May 2025, but under strict conditions that limited its role to specific administrative tasks, including drafting documents and summarizing general reports, with a categorical ban on entering any operational, confidential, or personal information.
Investigation timeline and professional standards
Emily Barry, Director of Communications at the IOPC, explained that the allegations raised are extremely serious and require an independent and comprehensive examination. Investigators are currently focusing on auditing the officer's full usage records of the tool during the period extending from May 2025 to June 2026 to verify whether breaches of the approved professional standards in the British police force occurred, noting that the officer has received a notice of gross misconduct without a final judgment being issued so far.
AI governance challenges in law enforcement
This incident highlights data privacy issues and the risks of AI governance within security and sensitive institutions. Despite the accelerating integration of generative tools to alleviate administrative and paperwork burdens—such as some police departments in the United States relying on Axon's "Draft One" platform to create incident reports from body camera recordings—the risks of hallucination and data leaks remain present.
The re-entry of sensitive data into cloud models recalls previous incidents, such as Samsung employees leaking software code and confidential meetings via "ChatGPT" in 2023. The current case in Britain emphasizes the critical difference between automation that assists office work and delegating AI tools to tasks involving human judgment and life-altering decision-making within the judicial and security systems.