Automated Facial Recognition Surveillance

Chi Onwurah Excerpts

Monday 27th January 2020 Commons Chamber

[Chi Onwurah (Newcastle upon Tyne Central) (Lab)]

Like all artificial intelligence, and unlike the spotter in a football crowd that the Minister cites, facial recognition technology automates the prejudices of those who design it and the limitations of the data on which it has trained. If it is not diverse by design, it will be unequal by outcome, so what minimum standards is he placing on this technology before it is rolled out?

[Kit Malthouse]

The hon. Lady is quite right to raise what has been a concern in the media, but none of the evidence from the trials thus far—[Interruption.] Okay, the concern has been elsewhere as well. However, none of the evidence in trials thus far is pointing to that disproportionality. One of the key things that the Met will be doing, however, is that, after every deployment—[Interruption.] Madam Deputy Speaker, I am trying to answer the hon. Lady’s question, but she is still barracking me from a seated position. I would like, if possible, to explain it. I understand that it is a very sensitive issue, but we are, nevertheless, dealing with very serious crime and this may help the police in apprehending those people. Frankly, if the police were seeking to apprehend the killer of my child, I would want them to consider using this technology. We owe it to people to make the police as effective as possible. However, the Metropolitan police will be publishing the results of every deployment on their website. The democratic scrutiny will be exposed through the London Assembly and, indeed, I am sure, through this place. As the technology is rolled out and we consider what changes may be needed to the legal framework so that it operates in a position of confidence with the public, no doubt Members here will have their say.