A patchwork rather than a single framework
There is no single statute in the United Kingdom that comprehensively regulates police use of artificial intelligence. As of the most recent independent legal analysis, no express statutory regulation exists specifically governing the use of AI in criminal proceedings at all. Instead, police forces operate under a combination of general laws never written with AI in mind, chiefly the Police and Criminal Evidence Act 1984, the Equality Act 2010, and the Data Protection Act 2018 together with UK GDPR, supplemented by non-statutory guidance from bodies including the College of Policing, the National Police Chiefs' Council and the Crown Prosecution Service. The government has been explicit that it favours this approach. Ministers have described a pro innovation stance toward AI regulation generally, and senior police leaders have argued publicly that forces should not be expected to wait for Parliament to legislate before adopting new technology.
That view is contested. Civil liberties organisations including Liberty have pointed out that this same patchwork has allowed forces to introduce technologies including live facial recognition and predictive analytics with limited governance for years before any serious regulatory response, and one mapping exercise cited by Liberty identified more than 70 distinct AI tools already in use across police forces in England and Wales, illustrating how far deployment has already run ahead of any settled legal framework built specifically for it.
A new national centre for police AI
The government's most significant institutional response so far is PoliceAI, a National Centre for AI in Policing that began mobilising in April 2026 and formally launched in June 2026, hosted within the College of Policing before an eventual move into the planned National Police Service. Its stated purpose is to consolidate AI related activity that would otherwise be duplicated separately by each of the 43 forces in England and Wales into a single national team, running trials, evaluating tools and issuing guidance intended to apply consistently across policing rather than force by force. Alex Murray, previously the National Police Chiefs' Council's AI lead and author of an earlier national AI strategy and adoption playbook, became the centre's interim director.
PoliceAI's own launch material sets out an expansive remit: large scale pilots across up to ten forces to help officers triage and summarise digital evidence, a national response to AI enabled crime including deepfake intimate imagery through a dedicated Policing AI Threat Hub, and continued rollout of AI tools for redacting audio-visual evidence, which the government estimates could save a million hours of officer time annually if adopted across all 43 forces. The centre sits within a wider Police Reform White Paper, published in January 2026, that the government has described as the most ambitious redesign of policing in nearly two centuries, encompassing far more than AI governance alone. Not every voice inside policing has welcomed the pace this implies. The head of the College of Policing was reported to have described the intended approach to spreading validated AI tools across forces in strikingly urgent terms, comparing the desired speed of adoption to a wildfire, a comment that drew criticism from observers who argued it signalled enthusiasm for speed over caution.
Legislation moving in parallel
Alongside PoliceAI, specific legislative changes have also progressed. The Crime and Policing Act 2026 received royal assent on 29 April 2026 and covers a wide range of policing matters, including provisions addressing AI generated child sexual abuse material and an amendment extending Online Safety Act powers to previously unregulated AI services, intended to bring tools capable of generating illegal content within regulatory scope. Separately, and specifically concerning facial recognition and biometrics rather than policing AI in general, the government launched a public consultation in December 2025 on a dedicated legal framework for law enforcement use of these technologies, responding to years of criticism that facial recognition in particular has operated without a bespoke statutory basis. As of the most recent reporting, the government's formal response to that consultation had not yet been published.
What is not settled
Whether PoliceAI functions primarily as a genuine, independent check on how forces adopt and evaluate AI tools, or chiefly as an accelerant for adoption with evaluation as a secondary function, is a live and disputed question, not a settled one, and will depend heavily on how the centre actually operates once its pilots move from planning into practice. The consultation on a dedicated legal framework for biometrics and facial recognition remains open in the sense that no resulting legislation has yet been enacted, meaning the technology that has generated the most legal challenges to date, including the Bridges litigation against South Wales Police, still operates without the bespoke statutory footing campaigners have long argued it needs. More broadly, the tension between a national government keen to demonstrate that policing is modernising quickly and independent scrutiny bodies warning that oversight has consistently lagged behind deployment is, on the evidence gathered here, not close to being resolved.
Follow the coverage
PoliceAI News tracks every regulation & oversight story from across the English-speaking world as it breaks: new deployments, court rulings, parliamentary debates, academic research, and community responses. The feed refreshes every 30 minutes.
View Live Regulation & Oversight StoriesYou can also browse every regulation & oversight story in the feed, or explore every subject PoliceAI News covers.