West Midlands Police Copilot scandal exposes AI hallucinations in UK policing decisions

West Midlands Police AI Scandal Reveals Critical Governance Gaps

The UK Parliament's Home Affairs Committee has released a damning assessment of how West Midlands Police integrated AI in policing operations, specifically regarding the controversial November ban on Maccabi Tel Aviv supporters attending an Aston Villa fixture. The committee's findings reveal that Copilot, Microsoft's AI tool, generated inaccurate evidence about disorder at a fictional West Ham match that never occurred, yet this AI-generated hallucination made its way into an intelligence document that shaped the public order decision to bar fans 1

1

.

The Home Affairs Committee findings show the problem extended beyond the chatbot's fabrication itself. MPs emphasized that the real failure lay in how easily the false information was accepted without verification, pointing to significant due diligence failings within the force. The report states it should not have required two oral evidence sessions and a written correction to establish an accurate account, signaling broader issues with transparency and attention to detail 1

1

.

Leadership Accountability and the Departure of Craig Guildford

Former chief constable Craig Guildford initially told the committee that officers had not used AI to find the material, only to later correct the record in writing. While MPs stopped short of accusing Guildford of deliberately misleading Parliament, they noted he was not briefed before his January 6 evidence session that AI had generated the incorrect material, despite the AI use being disclosed internally by that point. The committee criticized Guildford for showing a "remarkable lack of professional curiosity" by failing to properly verify evidence before appearing, with MPs stating that getting facts wrong twice indicated systemic problems rather than isolated errors 1

1

.

Following criticism from Home Secretary Shabana Mahmood and mounting pressure, Guildford retired at 52. The acting chief constable subsequently switched Copilot off across West Midlands Police while investigators examined what had transpired 1

1

.

Inconsistencies in AI Policy Across UK Forces

Despite the wrongful ban of football fans and the subsequent scandal, Sky News revealed that at least 21 police forces across England continue using Copilot, exposing troubling inconsistencies in AI policy and lack of coordinated oversight 2

2

. Only eight forces responding to inquiries stated Copilot could not be used in investigations, including police in Scotland and Northern Ireland 2

2

.

The National Police Chief's Council told Sky News it "is confident that the potential benefits of using AI outweigh the risks posed, provided we remain committed and vigilant in using it correctly, responsibly and securely." However, their guidance to use Copilot "in the most appropriate way" leaves decisions to local forces, creating a fragmented approach to regulatory oversight 2

2

.

Greater Manchester Police, England's second-largest force, defended its use, stating it has "a robust AI policy in place to help promote the use of such technology to speed up processes and ensure officers have more time to be on the streets." West Yorkshire Police said staff receive "education and guidance on how to use it responsibly, which should avoid any issues" 2

2

.

Questions About Responsible Implementation of AI

West Midlands Police and Crime Commissioner Simon Foster expressed concern about how the force was utilizing AI beyond this particular operation, noting "significant concerns, shortcomings, and failures around ensuring there was a proper regulatory management." Foster emphasized the need to ensure AI use is lawful, reasonable, and ethically deployed with proper governance to prevent misuse and erroneous results 2

2

.

The controversy has exposed how senior officers were unclear about how AI tool generated inaccurate evidence, highlighting broader concerns about technology deployed as a time-saving measure despite inherent risk. Chris Todd, chair of the National Police Data and Analytics Board, insisted AI should be used "to support human decisions, not make them for us," while acknowledging efficiency benefits 2

2

.

Microsoft stated it "continuously evaluates" Copilot and urges companies to review their usage 2

2

.

Implications for AI in Policing Policy

The scandal arrives at a critical juncture for policymakers. Last month, the government published a white paper outlining plans to expand AI in policing, including £115 million over three years for a new National Centre for AI in Policing called Police.AI, initially focused on automating administrative work 1

1

. The timing raises questions about whether adequate safeguards exist before scaling AI deployment across law enforcement.

The incident serves as a stark reminder that while AI offers potential to reduce administrative burden and improve data analysis, its integration requires rigorous accountability mechanisms, clear governance frameworks, and human rights considerations. As forces balance operational efficiency with public trust, the West Midlands case underscores that transparency and verification protocols cannot be optional extras but must form the foundation of any AI deployment in policing operations.