British Police Consider Using Data to Predict Future Criminal Activity

Facial recognition technology is operated in Sydney, Australia, in this file photo. (Ian Waldie/Getty Images)

The Met Police want to use data to predict which men will commit violent crimes against women or girls, according to its most senior officer.

Speaking to the Exceptional Policing conference on Oct. 12, Met commissioner Sir Mark Rowley talked about wanting to use behavioural data to make predictions about future crimes which would stop would-be offenders before they attack.

He said that the Met is working to “build a city-wide data picture of men who we know prey on and commit abhorrent crimes against women and girls across London, which is more sophisticated than ever before.”

Rowley said, “sadly, we know it is many tens of thousands of men.”

“With thousands of people repeatedly identified as suspects, but where there is sometimes insufficient evidence to prosecute, but could we start to rank-order some of the most dangerous offenders now at large in our city?” he said.

Forecasting

Rowley said that he wanted the police to “go further” to see if the police “can build a clearer picture of future risk, forecasting, and interdicting men who will commit violent crimes against women or girls, based on previous behaviour as statistically-tested risk factors.”

Rowley is a former head of the Met’s Counter Terrorism Command but left Scotland Yard in 2018 and had been consulting on specialist security projects and working on technology and data.

Sir Mark Rowley takes an oath to King Charles III at New Scotland Yard in central London on Sep. 12, 2022 (PA).

He was appointed by the former Home Secretary Priti Patel as the new Commissioner of the Metropolitan Police, replacing Dame Cressida Dick who resigned in February.

The Met commissioner said he could not give “the full analysis today.”

However, he added that he could say that the “success rate of this approach is proving impressive.”

“We’re increasingly able to understand the likelihoods of who will commit some of society’s most serious crimes, and use that to inform our thinking about prevention,” said Rowley.

“Could it help us catch some of the worst offenders, to find the women and children most at risk as a priority and protect them? Could it also help our partners and other agencies best focus their efforts too?” he added.

In January, preliminary research into a two-year pilot programme run by Durham Constabulary with Cambridge University found that AI can predict reoffending.

The Harm Assessment Risk Tool (Hart) was tested to see if it was better at predicting reoffending rates than human custody sergeants.

Both the machine-learning program and human custody officers assessed whether offenders brought into custody were at high, medium, or low risk of reoffending.

Hart’s algorithm did better than its human counterparts, with an accuracy rate of 53.8 percent, compared to a 52.2 percent accuracy rate for custody officers.

Durham Constabulary claimed that it stopped using Hart in September 2020.

Welsh police have trialed a futuristic technology that will identify wanted individuals almost instantly.

A Met Office spokesman told The Epoch Times by email: “We don’t have any further details to share at this time.”

Chris Summers contributed to this report.