根据英国警务行为独立办公室（IOPC）的信息，这名来自德比郡警察局的高级警官正面临涉嫌严重失职的调查。调查于今年8月启动，起因是德比郡警察局在完成内部审查后，于6月23日主动向监管机构提交情况报告。
调查内容不仅涉及利用人工智能检查拼写和语法等基础办公任务。相关指控称，该警官涉嫌使用微软Copilot参与制定政策决定、评估及管理风险，并在刑事案件调查过程中利用AI为决策提供理由和依据。此外，他还被指将涉及案件的敏感行动信息以及个人资料输入系统之中。
事实上，德比郡警察局早在2025年5月便已批准警员使用微软Copilot，以提升工作效率。授权范围主要包括撰写报告、整理文件以及信息摘要等行政事务。然而，警方同时明确规定，任何行动数据、敏感信息和个人资料均不得输入AI工具。
英国警务行为独立办公室参与事务主管艾米莉·巴里表示，这些指控性质严重，因此必须接受全面且独立的调查。调查人员将重点审查该警官在2025年5月至2026年6月期间对Copilot的使用情况，并判断其行为是否违反警务专业标准。
目前，这名警官已经收到涉嫌严重失职调查通知，但案件尚未作出最终认定，调查仍在进行之中。
事件也反映出人工智能正在逐步进入执法机构日常工作流程。此前，美国部分警察部门已经开始采用Axon推出的Draft One系统，该工具能够根据随身执法记录仪的录音自动生成警务报告草稿。虽然最终内容仍需警员审核批准，但外界持续担忧AI可能带来的错误信息以及对自动化工具的过度依赖。
类似问题今年也曾引发关注。美国犹他州一份由人工智能生成的警务报告曾闹出乌龙，系统因错误识别背景音频内容，竟在报告中声称一名警员“变成了一只青蛙”。幸运的是，该错误在报告正式提交前被人工发现并纠正。
与此同时，案件中涉及将敏感资料输入AI系统的问题，也让人联想到三星在2023年发生的事件。当时有员工将机密程序代码和会议内容提交给ChatGPT处理，从而引发企业对于生成式人工智能数据安全风险的广泛担忧。
随着越来越多公共机构开始部署人工智能助手，这起调查或将成为英国执法系统评估AI使用规范的重要案例，同时也可能影响未来人工智能在警务工作中的监管政策和应用边界。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
According to the Independent British Police Conduct Office (IOPC), the senior police officer from the Derby County Police Department is facing an investigation for serious misconduct. The investigation was launched in August this year as a result of the initiative of the Derby County Police to submit a status report to the supervisory authority on 23 June following the completion of an internal review.
The survey did not cover only basic office tasks such as the use of artificial intelligence to check spelling and grammar. Allegations were made that the officer was suspected of using Microsoft Copilot to participate in the formulation of policy decisions, the assessment and management of risks and the use of AI to justify and justify decisions in criminal investigations. In addition, he was accused of entering into the system sensitive operational information and personal data concerning the case.
Indeed, in May 2025, the Derby County Police Department approved the use of Microsoft Copilot by police officers to increase efficiency. The mandate mainly covers administrative matters such as the preparation of reports, documentation and information summary. At the same time, however, the police have made it clear that no operational data, sensitive information or personal data may be entered into the AI tool.
Emily Barry, Chief of Participation of the Independent Office of British Police Conduct, stated that these allegations were serious in nature and required a full and independent investigation. The investigators will focus on examining the police officer ' s use of Copilot between May 2025 and June 2026, and will determine whether he or she acted in violation of police professional standards.
Currently, the police officer has been notified of a suspected serious misconduct investigation, but the case has not yet been finalized and the investigation is still ongoing.
The incident also reflected the gradual introduction of artificial intelligence into the day-to-day workflow process of law enforcement agencies. Previously, parts of the United States police force had started using the Draft One system, launched by Axon, which enabled the automatic generation of draft police reports based on the recording of the accompanying law enforcement recorder. While the final content is still subject to police clearance, there are persistent concerns about the information that AI may cause and overreliance on automation tools.
Similar concerns were raised this year. An artificially intelligent police report from Utah in the United States, in which the system falsely identified background audio content, claimed that a police officer “was turned into a frog”. Fortunately, the error was manually detected and corrected prior to the official submission of the report.
At the same time, the case involved the input of sensitive information into the AI system and was reminiscent of the events of Samsung in 2023. At that time, confidential program codes and meeting content were submitted to ChatGPT for processing, raising widespread concerns about the security risks of generating artificial intelligence data.
As more and more public institutions begin to deploy artificial intelligence assistants, this investigation may be an important case for British law enforcement systems to assess AI usage norms, but it may also affect future regulation policies and application boundaries of artificial intelligence in policing.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.