Biometric Recognition Technologies in Schools

Lord Vaizey of Didcot Excerpts

Thursday 4th November 2021 Lords Chamber

[Lord Vaizey of Didcot (Con)]

My Lords, I congratulate my friend the noble Lord, Lord Clement-Jones—he is a friend—on calling this important debate. I salute his stamina in having participated in the previous debate and seamlessly moved on to lead this debate today. It is a mark of his global influence that, only yesterday, Facebook announced that it was withdrawing all of its facial recognition technology from its site. That technology has been around for some 10 years, and a billion people have consented to have Facebook use it on them, but the minute the noble Lord put down this debate, his colleague Nick Clegg clearly thought, “This is an issue I need to look into”. Who knows why Facebook really made this decision? One could take a noble view that it did so because it thinks that it is intrusive and unnecessary, or a cynical view: it is not making the company any money, so why put itself in harm’s way by continuing to use it? This is an important point.

I will talk more widely about the regulation of facial recognition technology, which is the issue that the noble Lord has put in front of us, with a particular focus on schools. It is a classic example of where technology has outpaced, as it were, the ability of regulators and policymakers to keep up to date. In many respects, facial recognition technology can have benign uses. I suspect that quite a lot of people in this Chamber open their phones using facial ID. We have our faces scanned when we move through the electronic gates at airports, when they are working. We can use facial recognition technology to organise our photos on our phones. More and more airlines are introducing facial recognition technology to allow you to check in seamlessly. So as a customer service to which you voluntary opt in, it is a good thing.

However, as the noble Lord pointed out, there are of course the inevitable and justifiable concerns about the creation of a big brother society—one that is made worse by the deployment of this technology while it is still in its relative infancy. It is one thing to debate its use in the UK but quite another to see how it is being used in a country such as China, where I gather that it is now an offence to leave your home without your phone. It was a reason why so many Hong Kong demonstrators wore masks.

One of the big problems with deploying facial recognition technology, apart from it being a bit of a word sandwich, is that it is in its infancy. We know it can be subject to bias. Frankly, it works more accurately for white men and white women. Amazon’s facial recognition tool incorrectly identified 28 Members of Congress as people who had been arrested, according to a test conducted by the American Civil Liberties Union. According to a paper published by the Massachusetts Institute of Technology and Stanford University, the technology struggles to identify people of colour and women. It has some rate of error even when operated in an unbiased way.

This has led to a decline in public support, which has dropped from about 50% to just over 40% in recent polls conducted in the USA. As my noble friend quite rightly highlighted, the debate is well under way. It is happening not just in this country but in the US, where the House Committee on Oversight and Reform has hosted hearings, and in individual US cities; for example, San Francisco’s Board of Supervisors passed a measure to ban the use of this software by law enforcement.

It is not just policymakers. Quite rightly, some significant companies—including Microsoft and Amazon, for example—have sought to get ahead of this debate and call on policymakers to regulate facial recognition. IBM published an important paper on facial recognition technology, which says that it should be used only where you have the ability to be given notice that it is being used and to consent. It called for export controls on facial recognition technology where it might be used for law enforcement or military purposes and said that law enforcement authorities should be mandated to disclose facial recognition technology and publish regular transparency reports. As my noble friend points out, the Information Commissioner’s Office has, as I understand it, been closely monitoring facial recognition technology trials, particularly those carried out by the British police, and is reviewing the regulations surrounding it.

It is important that this debate highlights that there remains a gap in how facial recognition technology is regulated and uncertainty over whether it falls in the regulations that apply to surveillance cameras and CCTV, and that we need a clear direction from the Government as to which bodies are responsible for overseeing the use of facial recognition technology—whether it is the ICO because it is a data protection issue, or the education authorities focusing on it as an education issue. It is also important that clear guidance is put out, so that people wanting to use facial recognition technology—as I say, there are many benign and quite convenient use cases for it—are aware of the basic principles they should adopt when they deploy this technology.