Site navigation

UK Policing AI and Automation Called Out Over Racial Bias

Elizabeth Greenberg

,

police ai
The scrutiny on police tactics follows days of riots driven by racism and Islamphobia that took place across the UK last week. 

AI and automation in UK policing is being called out to the UN for perpetuating structural racism within law enforcement and the wider country.

The joint report by Runnymede Trust and Amnesty International UK to the UN Committee on the Elimination of Racial Discrimination comes less than a week after riots by far-right nationalists broke out across the UK after the murder of three young girls was wrongly associated with a Muslim man.

The riots – driven by rising racism, Islamophobia, and xenophobia galvanised and exploited by rising far-right politicians – are only a symptom of structural racism perpetuated in the country.

“The racist and Islamophobic violence unfolding on the streets of the UK highlight the failures of successive governments to make progress on institutional racism,” Ilyas Nagdee, racial justice director at Amnesty UK said.

The report eviscerated the state of racism in the UK and the wider world, discussing the socioeconomic effects across different racial groups, and racisms intersections with gender, religion, and social mobility.

Notably, the report had a section specifically discussing the role of AI and automation systems in their perpetuation of racism, highlighting its use by police in England and Wales.

Since 2015, police forces in England and Wales have employed facial recognition technology, including the more controversial live facial recognition (LRF) technology. This technology is often used in large crowds in live time, so public consent is not gathered, and non-suspects may have their face scanned.

In 2020, the use of LRF by the South Wales Police was actually found in breach of the law, and is continually called out for racial bias, as well as its tendency to mis-identify black and minority ethnic people.

When the Met Police introduced AI LRF in 2020, the technology was actually found to be 81% inaccurate, sparking concerns.

The use of LFR was used nine times between 2020 to 2022, resulting in nine arrests or disposals, but this figure exploded to 96 occasions between 2023 and May 2024, resulting in 243 arrests.

LRF has notably been used in the policing of protests, including those in solidarity with Palestine, and the King’s coronation.

It was also revealed that the Home Office is already secretly using facial recognition in searches using its passport photo and immigration databases, sparking concerns around transparency.

Further, facial recognition has been found to be increasingly used in non-police settings, such as foodbanks and to combat shoplifting. According to the report, the Home Office’s actions appear in support of this, as it lobbied the Information Commissioner’s Office – the UK’s independent data protection regulator – to approve the use of private facial recognition software.

Coupling an intelligence system which is known to perpetual racial bias and mis-identify BME individuals with a policing force that has admitted to having a racial bias problem would logically compound in the exacerbation of structural racism in the UK.

The Metropolitan Police already faced scrutiny of their Gangs Matrix, which racially profiled predominantly young Black boys and men for being suspected gang members based on their music taste and social media activity – but similar systems empowered by AI, which is known to perpetuate human bias, could be even worse.

To combat this, the statement is calling for an inquiry into all police gang databases for racial bias and their overall effectiveness.

The paper therefore recommended the prohibition of all forms of predictive and profiling systems in law enforcement and criminal justice.


Recommended reading


The advocacy groups also call for transparency when police, migration, and other national agencies use ‘high-risk’ AI, such as facial recognition and decision-making AI.

The joint statement is also calling for legal limits prohibiting AI for uses posing an unacceptable risk to human rights.

“People of colour have faced a rowback on their civil and political rights over the past five years,” Alba Kapoor, head of policy at the Runnymede Trust, said.

“From the inhumane changes to our asylum system, to the introduction of legislation that restricts protest rights and ramps up harmful policing powers, it is time to shift the dial.”

Elizabeth Greenberg

Staff Writer

Latest News

AI

Nvidia Launches Open Secure AI Alliance for AI Safety and Security

AI Business Recruitment

Nearly a Quarter of Orgs Reducing Entry-level Hiring Due to AI Automation

Business

Scottish Businesses Turn to Self-funding as Growth Confidence Dips in H2

Data Finance

Payment Leaders are Struggling to Get Real-time Data