William - stock.adobe.com

Metropolitan Police chief warns against law updates amid substantial tech expansion

The Metropolitan Police is set to significantly expand its use of AI, drones and facial recognition in a bid to ‘regain the advantage’ over criminals, but commissioner Mark Rowley warns that progress could be held back by updated legislation and a lack of data integration in and between forces

Metropolitan Police commissioner Mark Rowley has set out the force’s plan to drastically upscale its use of artificial intelligence (AI), live facial recognition (LFR) and drone technologies, as part of a wider warning against police-specific technology legislation.

Speaking on 24 June, he presented the case for change in how the Met approaches technology deployment and integration, arguing that “policing today will fail if it is not allowed to keep pace with criminals” who are increasingly “organised, networked and digital”.

To “regain the advantage”, Rowley said the use of new technologies was imperative, helping police integrate and exploit the colossal amounts of data they already hold, save money and build trust.

As part of his warning, Rowley also announced a major roll-out of drone operations across the capital, the introduction of static live facial recognition cameras across central London, and the increased use of AI to analyse video evidence in criminal investigations.

Highlighting how the introduction of technologies such as fingerprinting or body-worn video were initially met with strong public concern when they were introduced, he said “the lesson from nearly 200 years of policing” is that while new tech often meets resistance, it becomes “indispensable” when used responsibility and proves its worth.

However, Rowley also took aim at campaign groups, which he accused of “slowing progress” with their calls for legislation on new technologies such as LFR.

“We already operate within a robust legislative framework … and we are overseen by at least five regulators and commissioners,” he said. “If every time a new capability emerges, or we find a new use case for data exploitation, we have to wait for new legislation before we use it, we will fail. We cannot legislate for every incremental development in technology. The pace is simply too fast, and the legislative process is too slow.”

Public sector procurement

In a similar vein, Rowley also called for changes to public sector procurement so that police are able to leverage new technologies at a much faster rate than the current system allows, which “delivers outcomes months, if not years, after they are needed”.

He added that a “radical overhaul” was needed to get new tools in the hands of police, noting that while the Home Office previously announced a new national policing model in January 2026, “we now need to deliver reform soon, or it will only be a matter of time before we face a disaster that a simpler model with integrated, advanced technology would have avoided.”

Part of this reform should include properly integrating currently disconnected data from across UK policing.

Rowley said that while the Met itself can gain real value in connecting the roughly 2,200 datasets it holds across more than 600 systems, analytics, automation and AI packages can also help connect it to data held by the other 42 forces in England and Wales if there are changes to the structure of policing.

“The current cost, complexity and delays in building modern technology 43 times over are prohibitive,” he said. “Without reform, new technology will have to be deployed into 43 different IT infrastructures. Criminals don’t respect boundaries, yet our current model makes national data exploitation nearly impossible.”

Rowley added that the public is more concerned about police failing to “join the dots” in its own data than the fact they used AI to do so.

Technological expansion details

On the potential of new technologies today, Rowley said AI-powered video analytics, for example, can help transform the collection and processing of information from London’s vast CCTV network – one of the biggest in the world.

“We are already seeing the impact in early use across 23 major investigations, with more than 16,000 hours of CCTV,” he said. “Video analytics reduced review time by 454 officer days, and that was just in 23 investigations. That is not just efficiency. That is officers freed up to investigate, to protect victims and to prevent further harm.”

On the expansion of the Met’s nine-strong drone fleet – which are already deployed to around 200 incidents per week, with average response times of roughly two minutes compared with the nine-minute average response of human officers – Rowley said the ambition is to have drone coverage in every London borough by this time next year.

“We are seeing them peruse offenders, locate missing people and reduce risk in real time, but this is only the start,” he said, adding that there is also an ambition to integrate the drone capabilities of London’s blue light services.

“This should not just mean sharing drones,” said Rowley. “It should mean sharing the underlying infrastructure that makes them effective, the airspace management, the launch sites, the connectivity, the data and control systems. Our aim should be a single, secure, coordinated infrastructure that allows police, fire and ambulance services to operate seamlessly as we protect the public.”

The Met’s planned major expansion of LFR – which he said has contributed to more than 2,000 arrests since 2024 – will see the force fix static LFR cameras on existing street furniture such as lampposts.

All of this, said Rowley, will help ensure greater trust in the Met. “Technology is not just reducing crime,” he added. “It is building trust. Officers and staff will remain at the heart of every decision. What this approach does is equip good people with better tools. It does not mean abandoning ethics – quite the opposite. It means strengthening them.

“There will be human oversight underpinned by clear values and accountability at every stage,” said Rowley.

However, the continued ramping up of LFR by police has been heavily criticised by multiple digital rights groupslocal councillors and London Assembly members

‘Robust’ legal frameworks?

In arguing that clear, consistent principles – underpinned by strong oversight, accountability and public transparency – are a more effective way of governing how police use technology than “a separate legal framework for each new capability”, Rowley highlighted the patchwork of legislation that already exists.

This specifically includes the Human Rights Act, the Investigatory Powers Act, the Data Protection Act, and the UK General Data Protection Regulation (GDPR), which Rowley described as “robust”.

The use of LFR by police in England and Wales – beginning with the Met’s deployment at Notting Hill Carnival in August 2016 – has already ramped up massively in recent years, but there has so far been minimal public debate or consultation.

Although the Home Office has claimed for years that there is already “comprehensive” legal framework in place, it formally launched a 10-week consultation on the use of LFR by UK police in December 2025, with a view to introducing a new legal framework for this and other biometric technologies.

At the start of the consultation – which the Home Office is still yet to formally respond to – the department said that although a “patchwork” legal framework for police facial recognition exists (including for the increasing use of the retrospective and “operator-initiated” versions of the technology), it does not give police themselves the confidence to “use it at significantly greater scale … nor does it consistently give the public the confidence that it will be used responsibly”.

It added that the current rules governing police LFR use are “complicated and difficult to understand”, and that an ordinary member of the public would be required to read four pieces of legislation, police national guidance documents and a range of detailed legal or data protection documents from individual forces to fully understand the basis for LFR use on their high streets.

Although the Met’s use of LFR was recently challenged in a judicial review, the case was ultimately dismissed by the High Court on the basis that the force’s LFR policy document provided sufficient constraints to prevent abuse and ensure compliance with human rights laws.

Read more about police technology

Rowley said the court “knocked that challenge down completely”, adding that it shows “there is a legal framework”.

Asked by Computer Weekly how the force is dealing with the ongoing illegal retention of millions of custody images by the Home Office – which can be used in LFR watchlists and which successive biometrics commissioners for England and Wales have flagged as a major issue – Rowley said “they’re not feeding into LFR”, adding that the force is “careful” to focus on people who are wanted to the police.

“If you’ve had a caution for the 14-year-old shoplifting, you’re a 30-year-old, and you get stopped because we’ve kept your image, and we’re using it in our facial recognition, that would be disgraceful,” he said. “And, of course, we’re not doing that. If we were doing anything like that, I’m sure the High Court would have absolutely castigated our approach.”

Computer Weekly was present for the judicial review, which revolved around whether the Met’s policy document allows too much discretion over where and who the technology targets. The unlawful retention of custody images and their potential use in LFR watchlists was not addressed by either side, nor mentioned in any of the submissions to the court.

While unrelated to the Met’s use of LFR, it was reported in February 2026 that Alvi Choudhury – a software engineer never convicted of a crime in his life – was arrested and detained off the back of an LFR match, despite having never been involved in the alleged burglary incident, because Thames Valley Police were still holding and using a five-year-old custody image of him.

This image was taken by the force after Choudhury was previously detained, but never charged or let alone convicted, over an altercation between two groups in 2021.

Read more on Big data analytics