Abstract:
A senior British police detective is under independent investigation for allegedly improperly using Microsoft's Copilot artificial intelligence assistant during a criminal investigation. The incident has once again raised concerns about the boundaries of artificial intelligence application in law enforcement and the security of sensitive data.
The senior police officer from Derbyshire Police is facing an investigation for alleged serious misconduct, according to the Independent Office for Police Conduct (IOPC). The investigation was launched in August this year after Derbyshire Police voluntarily submitted a situation report to the regulator on June 23 after completing an internal review.
The investigation goes beyond basic office tasks such as using artificial intelligence to check spelling and grammar. Related charges stated that the police officer was suspected of using Microsoft Copilot to participate in making policy decisions, assessing and managing risks, and using AI to provide reasons and basis for decision-making during the investigation of criminal cases. In addition, he is accused of entering sensitive operational information and personal data involved in the case into the system.
In fact, the Derbyshire Police Department has approved police officers to use Microsoft Copilot as early as May 2025 to improve work efficiency. The scope of authorization mainly includes administrative matters such as writing reports, organizing documents, and summarizing information. However, the police also clearly stipulate that no mobile data, sensitive information and personal data are allowed to be input into AI tools.
Emily Barry, director of engagement at the Independent Office for Police Conduct, said: "These allegations are of a serious nature and must be subject to a full and independent investigation. Investigators will focus on examining the officer's use of Copilot from May 2025 to June 2026 and determine whether his behavior violated professional standards of policing.
At present, the police officer has received a notice of investigation for suspected serious negligence, but the case has not been finalized and the investigation is still ongoing.
The incident also reflects that artificial intelligence is gradually entering the daily work processes of law enforcement agencies. Previously, some police departments in the United States have begun to adopt the Draft One system launched by Axon, which can automatically generate draft police reports based on the recordings of body-worn body-worn cameras. Although the final content still needs to be reviewed and approved by police officers, there are continuing concerns about misinformation that AI may bring and over-reliance on automated tools.
Similar issues have also attracted attention this year. A police report generated by artificial intelligence in Utah, USA, went wrong. The system misidentified the background audio content and actually claimed in the report that a police officer "turned into a frog." Fortunately, the error was manually discovered and corrected before the report was officially submitted.
At the same time, the case involves the input of sensitive data into the AI system, which is also reminiscent of the incident that occurred at Samsung in 2023. At that time, some employees submitted confidential program code and meeting content to ChatGPT for processing, which triggered widespread concerns among enterprises about the data security risks of generative artificial intelligence.
As more and more public agencies begin to deploy artificial intelligence assistants, this investigation may become an important case for the British law enforcement system to evaluate the norms of AI use. It may also affect future regulatory policies and application boundaries of artificial intelligence in police work.
Comments