🔗 Share this article

Law enforcement agencies across the United Kingdom effectively campaigned to use a facial recognition system known to be biased against females, youths, and individuals from minority ethnic backgrounds, following complaints that a more accurate version generated fewer investigative leads.

The Technology in Practice

UK forces utilize the police national database (PND) to conduct searches using historical face recognition. This procedure entails comparing a “probe image” of a suspect against a repository of over 19 million custody photos to find potential matches.

Admitted Bias

The UK interior ministry conceded last week that the technology was flawed. This admission followed a review by the National Physical Laboratory (NPL) determined it misidentified people of Black and Asian heritage and women at much greater frequency than white men. The ministry said it “took steps on the findings”.

“This raises the question of whether facial recognition only becomes effective if users tolerate biases in race and sex. Convenience is a poor argument for overriding basic freedoms.”

Known Issue

Official papers show that this discriminatory flaw has been known about for more than a year. Furthermore, police forces lobbied to reverse an earlier ruling that was designed to address the problem.

Senior officers were informed of the algorithmic discrimination in late 2024. The government-ordered NPL review concluded the system was more likely to produce false positives for images depicting women, Black people, and those under 40 years old.

A Policy U-Turn

In response, the national police leadership body ordered that the confidence threshold required for potential matches be raised to a point where the disparity was greatly diminished.

However, this directive was overturned the following month following complaints from police that the adjusted system was producing fewer “useful lines of inquiry”. Internal records indicate the higher threshold cut the proportion of queries resulting in possible identifications from 56% to a mere under 15%.

Severe Disparities

Although the Home Office and NPCC declined to specify what threshold is now in operation, the latest NPL study found the system could generate false positives for women of Black heritage almost 100 times more frequently than for Caucasian women at specific configurations.

The Home Office commented on these results: “Our evaluation found that in a limited set of circumstances the algorithm is more likely to incorrectly include some demographic groups in its match reports.”

Balancing Utility and Fairness

Outlining the effect of the temporary raise to the system's confidence threshold, the NPCC documents note: “This adjustment greatly lessens the impact of bias across legally safeguarded attributes of ethnicity, age and sex but had a substantially detrimental effect on police efficiency”. The documents further note that police units argued that “a once effective tactic returned outcomes of limited benefit”.

Wider Implementation Proposals

Meanwhile, the government has opened a two-and-a-half-month public review on its proposals to expand the use of facial recognition technology. Policing minister Sarah Jones has labeled the technology as the “most significant advance since genetic fingerprinting”.

Criticism from Advisors and Monitors

Abimbola Johnson, chair of the independent scrutiny and oversight board for the police race action plan, commented: “There was very little discussion in equality strategy sessions of the facial recognition rollout despite obvious cross-over with the strategy's goals.

“These revelations demonstrate once again that the pledges to combat discrimination policing has undertaken through the equality initiative are not being translated into broader operations. Our reports have warned that innovative tools are being implemented in a landscape where ethnic inequalities, inadequate oversight and faulty information gathering already persist.

“Any use of facial recognition must adhere to strict national standards, be subject to external review, and prove it reduces rather than compounds ethnic bias.”

Home Office Response

A Home Office spokesperson stated: “We treat the conclusions of the study seriously and we have already taken action. A updated software has been externally evaluated and procured, which has demonstrated no measurable discrimination. It will be trialled early next year and will be subject to evaluation.

“Our priority is protecting the public. This revolutionary tool will assist officers to put criminals and rapists behind bars. There is officer review in each stage of the process and no further action would be taken without specialist personnel carefully reviewing the results.”