UK to spend £15M on AI-powered crime mapping in knife violence crackdown
Home Office hopes tech will help cops target hotspots as ministers push to halve offenses
The British government is spending £15 million over the next three years to improve crime mapping in England and Wales, partly to allow more targeted policing of knife crime.
The Home Office says its national mapping tool, which divides England and Wales into 1.46 million hexagons of around 0.1 square kilometers, found that all the knife crimes recorded by police from April 2024 to March 2025 took place in fewer than 2.5 percent of the hexagons.
The government's funding body, UK Research and Innovation (UKRI), reckons it can deliver improved data visualization and analysis of police-recorded crime data, add other types of data, and use AI to improve pattern recognition.
"By seeing knife crime at this micro-geography, local partners can pinpoint the specific streets, times and drivers of crime and direct the right mix of policing, prevention and services to those places," the Home Office states in "Protecting Lives, Building Hope," a policy paper detailing plans aimed at halving knife crime.
Although 2.5 percent of England and Wales may sound small, it is also the total area covered by the Metropolitan Police, Greater Manchester Police, and West Midlands Police forces, which between them record 43 percent of knife crimes. However, examples of the maps in the paper show big differences in rates even in city centers.
A hexagon that includes Leicester Square and Trafalgar Square in central London recorded more than 45 knife crimes between April 2024 and March 2025, while another nearby that includes Long Acre in Covent Garden saw fewer than five. In Birmingham, three neighboring hexagons that include the Grand Central and Bullring shopping centers each saw 18 or 19 knife crimes in that period, while two adjoining ones recorded fewer than five.
- UK police force presses pause on live facial recognition after study finds racial bias
- Cops get more facial recognition vans as UK bets big on AI policing
- High Court to grill London cops over live facial recognition creep
- Metropolitan Police hails facial recognition tech after record year for arrests
The paper says the government spent £5 million last year on "hyperlocal knife crime pilots" using the data in 11 police force areas and plans to spend £26.25 million this year across 27 forces. It adds that interventions can include police patrols in locations and at times when knife crime often takes place, greater use of knife-detecting wands and arches, and more CCTV including for retrospective facial recognition.
It also mentions the Home Office's expansion of live facial recognition (LFR) from 10 to 50 vans across England and Wales, first announced in January.
Last month, Essex Police suspended its use of LFR after a Cambridge University report found that its system was more likely to identify Black people than those from other ethnic groups on a watchlist.
The Protecting Lives, Building Hope policy paper also says that UKRI will undertake research into technology to prevent knife crime by looking for behavior that suggests people intend to commit violence with knives.
"This could provide staff in places like shopping centres, train stations and late-night venues with clearer guidance on what to look for and could also help ensure Stop and Search powers are deployed more accurately," it states. ®
More about
More about
Narrower topics
- AIOps
- Cabinet Office
- Competition and Markets Authority
- Computer Misuse Act
- CSAM
- DCMS
- DeepSeek
- GCHQ
- Gemini
- Google AI
- GPT-3
- GPT-4
- HMRC
- Home Office
- ICO
- Interpol
- Large Language Model
- Machine Learning
- MCubed
- NCSC
- Neural Networks
- NHS
- NLP
- Retrieval Augmented Generation
- RPA
- Star Wars
- Tensor Processing Unit
- TOPS
Broader topics
More about
More about
More about
Narrower topics
- AIOps
- Cabinet Office
- Competition and Markets Authority
- Computer Misuse Act
- CSAM
- DCMS
- DeepSeek
- GCHQ
- Gemini
- Google AI
- GPT-3
- GPT-4
- HMRC
- Home Office
- ICO
- Interpol
- Large Language Model
- Machine Learning
- MCubed
- NCSC
- Neural Networks
- NHS
- NLP
- Retrieval Augmented Generation
- RPA
- Star Wars
- Tensor Processing Unit
- TOPS



