Lord Boateng
Main Page: Lord Boateng (Labour - Life peer)Department Debates - View all Lord Boateng's debates with the Home Office
Police: Facial Recognition Technology
Lord Boateng Excerpts(3 days, 2 hours ago)
Lords ChamberRead Full debate Read Hansard Text Watch Debate Read Debate Ministerial Extracts
Lord Hanson of Flint (Lab)
I tried to answer my noble friend’s initial Question as best as I could. Procurement is another issue we are looking at. In the Government’s forward look to policing, we are considering what areas of work we can bring in centrally in terms of the guidance and support for the 43 police forces currently operating. Again, without pre-empting my right honourable friend the Home Secretary’s review, one possibility is giving greater guidance on procurement and issues such as facial recognition technology and other forms of preventive activity by police forces.
Lord Boateng (Lab)
My Lords, a US Government study suggests that facial recognition algorithms are far less accurate in identifying African-American and Asian faces than Caucasian faces, and that African and Asian women are 10 to 100 times more likely to be misidentified than Caucasian ones. The study identified 99 developers, including Intel, Microsoft, Toshiba and the Chinese firms Tencent and DiDi Chuxing, as potential problems in this area of procurement. What research are the UK Government going to commission on this, and how are these firms to be treated for the purposes of procurement by police forces in this country?
Lord Hanson of Flint (Lab)
My noble friend touches on important issues and again, I refer to the point I made earlier to the noble Baroness. A survey of the existing use of facial recognition technology estimated that there was no discrepancy between gender and race. My noble friend shows slight dissatisfaction with that potential outcome, and I say to him that those are the very factors we want to look at in the guidance my right honourable friend is considering bringing forward. Self-evidently, if we are going to use facial recognition technology, it needs to be accurate, regulated, proportionate, intelligence-led and organised in a way that does not discriminate against sex, race or any other characteristic.