Should the police ban the use of AI?

February 24, 2026·13 min·1 clip

West Midlands police banned Israeli football fans using AI-generated evidence that cited a football match that never happened.

The episode examines the use of AI by UK police forces, centred on a specific scandal involving West Midlands police and the Maccabi Tel Aviv fan ban. Sports correspondent Rob Harris explains that Aston Villa were drawn to play Maccabi Tel Aviv in the Europa League and that the match carried political significance given the Gaza conflict and a large pro-Palestinian Muslim population near Villa Park. West Midlands police decided in October to ban Maccabi fans entirely, a decision condemned by Prime Minister Keir Starmer. The decision was made by Birmingham City Council's safety advisory group on police recommendation, drawing heavily on events in Amsterdam in 2024 when Maccabi played Ajax and there was street unrest. Harris describes the investigative process: police would not publicly explain their decision, so documents were leaked. A leaked intelligence document from late October contained a reference to a previous Maccabi match against West Ham in 2023. The match had never been played — it was an AI hallucination. This was produced by Microsoft Copilot. Police initially denied using AI, claiming it was a Google search, before admitting the truth when recalled to the Home Affairs Select Committee. The chief constable, Craig Guilford, was forced into early retirement — Harris notes a chief constable leaving under such circumstances had not happened in decades. The Home Affairs Select Committee report identified fresh evidence of AI use as central to the intelligence document, and Amsterdam police told Harris they did not recognise some of the information West Midlands had cited. West Midlands subsequently switched off Copilot, which prompted Harris to survey every force in England. He found at least 21 are still using it in some form, with Police Scotland and Northern Ireland not using it. There is no national policy from the National Police Chiefs Council, which says only that AI benefits outweigh risks. Harris tested both the free Copilot Chat and the paid enterprise version with the same question about Maccabi fans. The enterprise version said the hooligan label was inaccurate; the free version weighted the answer more towards yes, drawing on different and arguably less neutral sources. Microsoft responded by saying they continuously evaluate their services and encourage organisations to use their own governance and review processes. Harris concludes that the fundamental problem is the absence of joined-up national thinking, with forces individually deciding whether and how to use AI — a situation he argues undermines public confidence in policing.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from This Is Why