Ethnic minority people are four times more likely than their white counterparts to be chosen for some police facial scans, according to official data, sparking warnings of racial bias.
uncertain
people → choose → bias
Internal figures released by South Wales police revealed that 24% of people selected by police for facial recognition checks were from ethnic minorities, despite this group making up only 7% of the area’s total population.
asserted
group → release → population
The force polices Wales’s largest cities – Cardiff and Swansea – and has pioneered several versions of AI-enabled facial recognition technology.
asserted
force → police → technology
The latest figures relate to police officers capturing a person’s face with a smartphone and running the image through an algorithm called “Neoface” to match it against custody mugshots and other face databases.
asserted
officers → relate → mugshots
Most scans are made because people are suspected of crimes.
asserted
people → make → crimes
Civil liberties campaigners called the ethnicity figures “alarming”.
asserted
campaigners → call → figures
The force said they raised “concerns about disproportionality” and that it was reviewing its approach.
asserted
it → say → approach
The data relates to a full year of use in 2025 and was obtained under the Freedom of Information Act and shared with the Guardian.
asserted
data → relate → Guardian
Even when compared with arrest data for the police area, in which 11% of people arrested are from ethnic minorities, the face-scan figures appear disproportionate.
asserted
figures → compare → minorities
South Wales is one of the first UK constabularies to routinely use this “officer-initiated facial recognition” (OIFR) technology, which aims to identify suspects, wanted criminals and missing people who decline to give their details when challenged as well as people who are found unconscious, uncommunicative or dead.
asserted
who → use → details
A similar system is being trialled by the Metropolitan police in London and has also been tested in Liverpool.
asserted
system → trialle → Liverpool
Critics have called facial recognition tools “Orwellian” and there have been repeated concerns about algorithmic bias putting ethnic minorities at greater risk of wrongful arrest.
asserted
bias → call → arrest
Last week, the police watchdog warned that innocent people, especially from ethnic minorities, remained at risk of wrongful arrest if police treated inaccurate retrospective facial recognition data as evidence.
asserted
police → warn → evidence
But advocates – including the home secretary, Shabana Mahmood – believe facial recognition is the biggest breakthrough for catching criminals since fingerprints and DNA matching.
asserted
recognition → include → fingerprints
Other versions of the technology include live facial recognition cameras directed from marked vans and cameras fixed to lamp-posts, like those piloted in Croydon between October 2025 and March 2026, which led to 173 arrests, including of rape, kidnap and sexual assault suspects.
asserted
which → include → suspects
“This data shows that intrusive biometric tools are being used to disproportionately target minority communities,” said Jake Hurfurt, the head of research and investigations at Big Brother Watch, a civil liberties campaign that obtained the figures.
asserted
that → show → figures
“It is alarming that non-white people have been subject to face scans at a much higher rate than would be expected either by the region’s demography or the makeup of people arrested in the region.”
asserted
people → expect → region
Aliya Mohammed, the chief executive of Race Equality First, a Cardiff-based anti-discrimination organisation, said the figures raised “serious concerns about racial disproportionality”.
asserted
figures → base → disproportionality
“Facial recognition is not neutral because it uses AI,” she said.
asserted
she → use → AI
“If existing biases influence who is stopped and scanned, technology can reinforce and even amplify inequality.”
asserted
technology → influence → inequality
Alvi Choudhury, 26, who was arrested in February for a burglary in a city he had never visited after police face-scanning software confused him with another person of south Asian heritage on a custody database, said it was “very concerning” that ethnic minorities were being disproportionately scanned.
asserted
minorities → arrest → database
“The police are under pressure to solve crimes and they are over-reliant on these AI tools,” he said.
asserted
he → solve → tools
“They shouldn’t be arresting people based on the AI tools alone.
asserted
They → arrest → tools
He remained worried AI systems were not trained on enough ethnic minority faces to ensure the number of false positives was no greater than for white people.
asserted
number → remain → people
According to the data for 2025, South Wales police officers initiated face scans of 257 adults and 14 children, including two under-10s.
uncertain
officers → accord → under-10s
More than half were suspected of a crime, being unlawfully at large or breaching bail conditions.
asserted
half → suspect → conditions
The next most common reasons were because officers believed the person posed a risk of harm to themselves or others – such as stalkers or sex offenders.
asserted
person → believe → stalkers
In about one in 10 cases, a scan was initiated for a suspected missing person while the technology was used to identify a dead person about once a fortnight.
asserted
technology → initiate → person
In many of the cases, the scan failed to deliver a match or the subject agreed to share their identity.
asserted
subject → fail → identity
Where a match was made, arrests followed in 48% of cases and no further action resulted in a third of cases.
asserted
action → make → cases
A spokesperson for South Wales police said: “We recognise the figures raise concerns about disproportionality.
asserted
figures → say → disproportionality
The number of OIFR searches remains relatively small, and the use of the technology is in its infancy.
asserted
use → remain → infancy
That said, each deployment is based on the individual circumstances that officers are dealing with at that time and is subject to supervisory and force-wide scrutiny.”
asserted
officers → say → scrutiny
They said the force would “continue to review our approach through our established governance processes, which also include input from independent advisory groups and external scrutiny panels” and added: “Where opportunities for learning or improvement are identified, we will take appropriate action.”
asserted
we → say → action
Earlier this month, the Home Office announced a new facial recognition watchdog with the power to stop police forces using the technology if they fail to follow rules and new laws in the police reform bill, including a duty for the use of face scans to be intelligence-led and proportionate.
asserted
use → announce → scans
The Home Office said decisions on the operational use of facial recognition were matters for the police, but they must comply with equality, human rights and data protection laws and its use must be necessary, proportionate and fair.
asserted
use → say → laws