At a Glance
- West Midlands police used Microsoft Copilot to write an intelligence report that wrongly banned Israeli fans from a match
- The AI invented a fictional game between Maccabi Tel Aviv and West Ham that never happened
- Chief constable Craig Guildford confessed after weeks of denials
- Why it matters: AI errors are now dictating public-safety decisions with real-world consequences
A UK police force has admitted its officers relied on Microsoft’s AI assistant Copilot to produce a flawed intelligence report that led to Israeli soccer fans being barred from attending a match in Birmingham. The revelation follows weeks of denials and has sparked fresh concern over the unchecked use of generative AI in critical public-safety decisions.
The Phantom Match That Never Took Place
Late last year, Maccabi Tel Aviv traveled to Birmingham to face Aston Villa in a European fixture. Ahead of kickoff, the local safety committee banned the Israeli club’s supporters from entering the stadium, citing a West Midlands police assessment that labeled the game high-risk for hooliganism.
The decision rested on a single intelligence report. That document, it has now emerged, was assembled with the help of Microsoft Copilot and included a wholly fabricated match: a supposed clash between Maccabi Tel Aviv and West Ham. The two teams have never met, and on the date referenced, West Ham were actually playing Olympiacos in Greece, according to the U.K. Parliament’s Home Affairs Committee.
Confession After Denials
In a letter to the committee, West Midlands chief constable Craig Guildford wrote: “On Friday afternoon I became aware that the erroneous result concerning the West Ham v Maccabi Tel Aviv match arose as result of a use of Microsoft Co Pilot.”
The admission reverses weeks of official denials that AI had any role in compiling the report. Guildford’s subordinates had failed to verify Copilot’s output before submitting the document to the safety committee, leading to the fan ban and subsequent political fallout.
A Pattern of AI Failures
The West Midlands episode is the latest in a string of high-profile hallucinations produced by enterprise AI tools. In October, consulting giant Deloitte refunded part of a $290,000 contract after delivering the Australian government a report filled with fictitious academic papers and court judgments generated by AI.
Despite these failures, American tech giants continue to push rapid adoption. Microsoft has made Copilot mandatory for its own workforce and markets the assistant as a productivity booster. The U.S. House of Representatives began using Copilot late last year, and companies across the American corporate landscape have deployed it widely.
Nvidia CEO Jensen Huang, speaking at an October briefing, said: “America needs to be the most aggressive in adopting AI technology of any country in the world, bar none, and that is an imperative. We have to encourage every single company, every single student, to use AI.” Huang has since dismissed discussions of AI risk as “hurtful” and “not helpful to society.”
Key Takeaways
- West Midlands police have admitted using Microsoft Copilot to fabricate intelligence
- The AI invented a non-existent soccer match, leading to Israeli fans being banned
- Chief constable Craig Guildford reversed weeks of denials in a letter to Parliament
- Previous AI blunders include a $290,000 Deloitte report packed with fake research
- Major tech firms and government bodies continue to expand AI use despite documented errors
