Met Police chief backs legislative framework for police tech

LFR already operational

Despite Dick calling for a legislative framework to govern the police’s use of algorithmic technologies, the Metropolitan Police began deploying LFR operationally for the first time in February 2020 in the absence of national guidance and in spite of earlier calls for it.

In October 2019, for example, following a 17-month investigation into police forces’ use of LFR, the Information Commissioner’s Office (ICO) recommended that the government introduce a statutory and binding code of practice on its deployment.

“I would argue that most areas in which we are already using modern technology are largely uncontroversial to the public,” said Dick, before attempting to dispel “some current and apparently pervasive myths” about the Metropolitan Police’s use of LFR.

Dick claimed there is a “very strong” legal basis for LFR use by police and that human officers will always make the final decision.

On the MPS website, the force lists the laws and legislation it claims enables it to use LFR, which includes the Human Rights Act 1998 and the Data Protection Act 2018 among others.

However, according to a July 2019 report from the Human Rights, Big Data & Technology Project, based at the University of Essex Human Rights Centre, it is highly possible that police deployment of LFR may be held unlawful if challenged in court because “no explicit legal basis exists authorising” its use.

It concludes the “implicit legal authorisation claimed by the MPS… is likely inadequate when compared with the ‘in accordance with the law’ requirement established under human rights law”.

The report, which marks the first independent review of the MPS LFR trials, also highlighted a discernible “presumption to intervene”, meaning it was standard practice for officers to engage a matched individual.

Dick also claimed that the technology used by the MPS is proven not to have an “ethnic bias”, adding the only bias is that “it is slightly harder to identify a wanted women than a wanted man.”

This is despite the fact the MPS’ facial recognition software, which is provided by Japan’s NEC Corporation, has never undergone any demographic testing.

On top of this, numerous studies exist pointing to a racial bias in similar facial recognition software. In the UK specifically, black people are three times more likely to be arrested than white people, according to the government’s most recent statistics, but no more likely to be convicted.

The same is true of people with mixed ethnicity, who are more than twice as likely to be arrested than white people.

In March 2019, the Science and Technology Committee heard there are more than 23 million custody images on the Police National Database (PND), regardless of whether the person was subsequently convicted or not.

This suggest that, given these custody images are used to create LFR “watchlists”, innocent people of colour are far more likely to be scanned, and subsequently engaged by police, than innocent white people.

In May 2019, the BBC reported that the MPS missed at least three chances to assess how well the systems deal with ethnicity over the past five years.