PARLIAMENTARY DEBATE
Facial Recognition: Police Use - 13 November 2024 (Commons/Westminster Hall)
Debate Detail
That this House has considered police use of live facial recognition technology.
It is a pleasure to serve under your chairmanship, Dame Siobhain. I am grateful for this opportunity to debate the police’s use of live facial recognition technology. I have to say that this debate is somewhat overdue.
Any fan of Hollywood movies would think that the use of facial recognition technology is widespread, as in “The Bourne Ultimatum” and “Spooks”, and that it is commonplace for MI5 and the CIA to tap into CCTV cameras across London. I do not believe that is correct— I hope it is not—but police forces are using facial recognition technology more and more. It was first used in 2017, and it is now commonly used by the Metropolitan police, South Wales police and now my own police force in Essex, which purchased two vans in August and use it regularly.
On 4 October, I accompanied police officers on a deployment in Chelmsford High Street, who were hugely helpful in explaining to me exactly how they use the technology and, importantly, what controls are in place. They told me that they had a watch list of 639 individuals who had been approved by the superintendent and were wanted for questioning in relation to offences such as violence against the person. They included people with outstanding warrants, suspects linked to county lines, suspected shoplifters in that particular part of the county, and those with a sexual harm prevention order.
In the course of the 30 minutes or so that I spent with those officers, they recorded 1,500 faces of people who passed by. The officers assured me that those images were matched against the watch list to see whether they registered a positive, and if they did not they were deleted in less than half a second. During the time I was there, there were approximately 10 positives, which led to a conversation: a police officer would go and have a polite exchange to find out why the person had registered positive, and they were checked against the Police National Computer or Athena. That morning, that led to two arrests.
The chief constable of Essex has written to me and colleagues to emphasise the effectiveness of the technology and its importance to that force. He told me that they had so far had 25 deployments across Essex, resulting in 26 arrests and 26 other positive disposals. He said:
“This cutting-edge technology has enabled us to keep the public safe, and can save time and effort of our front-line, allowing them to do other work to protect and support the community.”
I attended a meeting with Baroness Chakrabarti, along with my right hon. Friend the Member for Goole and Pocklington, where Shaun Thompson, an anti-knife community worker, spoke to us. He had been held by the police for 30 minutes and forced to provide all sorts of identity documents, as a result of a false positive. On the extent to which it is occurring and whether racial bias is involved, there is some evidence that that is the case. That makes it all the more important that we provide assurances.
We have heard from several campaign organisations that are concerned about the use. They vary in the extent to which they believe it is a legitimate technology. Big Brother Watch has described live facial recognition technology as
“constant generalised surveillance”
and has said that it is
“indiscriminately subjecting members of the public to mass identity checks”
which undermines the presumption of innocence.
Liberty has gone further, saying:
“Creating law to govern police and private company use…will not solve the human rights concerns or the tech’s inbuilt discrimination…The only solution is to ban it.”
I do not agree with that, because I think there is clear evidence that it has a real benefit in helping the police apprehend people who are wanted for serious offences, but one of my major concerns is the lack of any clarity in law about how it should be used.
I am grateful to the Library, which has provided advice on that point. It says:
“There is no dedicated legislation in the UK on the use of facial recognition technologies.”
Instead, its use is governed by common law and by an interpretation of the Police and Criminal Evidence Act 1984, although that Act does not mention live facial recognition technology, and some case law, such as the Bridges case. Even in the Bridges case, the Court of Appeal found that
“The current policies do not sufficiently set out the terms on which discretionary powers can be exercised by the police and for that reason do not have the necessary quality of law.”
The advice that came back following consultation with the Information Commissioner’s Office was that there is no blanket approval by the ICO for the use of LFR technology. Essentially, it should be judged on a case-by-case basis, but the ICO had expectations that data protection and privacy should be respected. It went on to say that the use of LFR can be highly intrusive and future uses of the technology may require updates, but that the ICO is monitoring it closely. That is only partially reassuring. Essentially, the ICO recognises that breaches of data protection could be possible, and is monitoring it, but there is no clear guideline to assist the police or anybody else with precisely how it should be used.
I am grateful to legal consultants Handley Gill, who wrote to me yesterday and who are involved in advising a number of people about the legality of the technology. They said that
“it is undesirable for individual Chief Officers and PCCs to have to engage in the wide ranging review and preparation of the necessary documentation, and that a move toward a common national approach (and choice of technology provider) would secure efficiencies and also enable closer monitoring…to ensure their efficacy and lawfulness.”
Although we are no longer bound by European Union law, the EU has brought in much more stringent controls than exist here.
“the AIA 2024 prohibits the use of ‘real-time remote biometric identification systems’ (such as LFR) in publicly accessible spaces for the purposes of law enforcement, unless such use is ‘strictly necessary’ for one of the following objectives”.
The list it provides includes the search for specific victims of abduction or trafficking; missing persons; the prevention of a substantial and imminent threat to life; the prevention of a genuine threat of a terrorist attack; or the localisation of a person suspected of having committed a criminal offence.
In Europe, the controls are strong, but in this country it is left largely to police officers to interpret the law and be reasonably confident. However, legal challenges are under way. Shaun Thompson, whom I met, is seeking judicial review of the police’s actions and the campaign organisations are also looking at legal challenges. There is a real need for clarity. Certainly, the sergeant of Essex police who is in charge of deployment told me that, in his view, it would be really helpful for the police to have clear guidelines. They would then not have to make those difficult decisions and could potentially satisfy a court that the use was proportionate and justified.
As far as I am aware, this matter has not been debated by Parliament before, and it should have been because there is a real need to seek clarity in the law. This may sound like science fiction, but ultimately there is a risk that it becomes possible for every CCTV camera in the country to be linked up, and there could be a watchlist of not 600 but millions of people. Concerns have been expressed by organisations such as Big Brother Watch—in this particular instance, that organisation could be well named—and I do not think any Member would wish to go down that route. I think most people recognise that there is some value in the technology, but there is a need for clarity. I am grateful to the shadow Home Secretary and particularly the Minister for Policing for coming to contribute, and I look forward to what they have to say.
Live facial recognition changes one of the cornerstones of our democracy: an individual is innocent until proven guilty. With this technology, if the machine says an individual is guilty because they have been identified using live facial recognition, they then have to prove their innocence. That is a huge change in our democracy that nobody has consented to. We have not consented to it in this place, and as we police by consent as a society, that should really worry us all.
On the question of changing the burden of proof or undermining the concept that someone is innocent until proven guilty, the technology absolutely does not change that. What it does is give the police a reason to stop somebody and check their identity to see whether they are the person wanted for a criminal offence. It certainly does not provide evidence on which a conviction might be secured. In fact, it is no different from the police stopping someone because they are suspicious of them, and it is a lot more accurate than stop and search, about which I am sure the hon. Lady has views. It is simply a tool to enable the police to stop somebody and check their identity to see whether they are the person who is wanted. It certainly does not undermine the very important principle that a person is innocent until proven guilty.
The thing is that we know very little about the software or what is in the black box that is developed by these systems. What we can look at is the outcome, and we know that the outcome does not identify very well black women’s faces, especially, and black and Asian people. There is a lower identification threshold for those people, so that is a concern.
It is also really interesting that even when LFR is set at 0.6, a police super-spotter is more accurate. We have specialist police officers who spot people very quickly, and they are more accurate than this system, so it becomes the case that a police service will try to prove that the system it has bought is value for money. We can imagine a police officer not getting many hits with LFR at 0.6 and lowering that to 0.5 so that they can get more hits, which in turn means that more people are misidentified, so there should be regulation around this issue.
Taking away somebody’s liberty is one of the most serious things we can do in society, so we need to think very carefully if we are going to introduce something that accelerates that. It is good that for the first time we are having the debate on this issue. As the right hon. Member for Maldon said, the EU permits LFR only where there is prior judicial authorisation and in cases in which the police need to locate a missing person, for instance. That is something we need to consider.
I want to say this: I like technology. I am very much into our civil liberties. We need to protect our digital rights as human beings and individuals. I love technology— I used to be a coder—but we should not rush to do things because people get excited. There are really four people in the debate on this issue. It reminds me of four of my mates when we go out clubbing. Bear with me. We have the person who will stay at home because they are not bothered—they do not care—and we have the people who do not care about this issue: “It is going to happen; let it happen.” We have the person who will come, but they are a bit moany. They do not really like the music, but they will come anyway because they do not want to miss out.
We then have the person who is completely drunk on it all: “Give it to me. I’ll take everything.” There are people who just love anything to do with technology and will say, “Look, let’s just throw it all in the mix and it’ll all be fine.” And there is me. I am the person who likes the music and the food, but I need to keep sober to make sure everyone gets home safely. In this debate about AI, we need to be sober to make sure that everybody gets home safely and that when we roll out AI, we do so in a way that is fair and compassionate and in line with our values as British citizens.
I have to make a confession to the House: I am not technically minded. I can just about use my phone for text messages; I cannot do much else with it. When it comes to TikTok, Facebook, X and all those other things, I am not even sure what they all are. The fact is that my staff do all that, so anything that people see on there from me is because of them. I okay it, but they put it out.
But even if I am not technically minded, I understand the necessity to have technological advances in place and that they can also be used to benefit our police and criminal justice system. I am fully in support of advancements where there is necessity and reason for them, but the hon. Member for Brent East (Dawn Butler) was right to identify some problems with the system. So my contribution will be in favour of facial recognition technology, but also focused on the need to have a system that does not infringe on human rights.
It was incredibly helpful to hear the comments of the right hon. Member for Maldon, and about how he was able to join police forces to see how live facial recognition works. I understand that was the 13th use of the technology by Essex police, with it having been deployed previously in Harlow, Southend and Clacton. Essentially, the equipment works by scanning the faces of all individuals seen by a camera and comparing them to a predetermined watchlist.
As my hon. Friend the Member for East Londonderry (Mr Campbell) said, safety is paramount—that is the critical reason for using the technology. I speak on human rights issues all the time, as many present will know. I want to make sure that when we have technology in place, human rights are not abused or disenfranchised, and that people do not feel threatened. Innocent people should never feel threatened, of course, but there are those who have concerns. The technology has already proven itself and led to a number of arrests of people wanted for serious offences such as sexual abuse, domestic violence, aggravated burglary and shoplifting.
I will make a quick comment about the Police Service of Northern Ireland and what we are doing back home. A freedom of information request was submitted to the PSNI in late 2022, and it was concluded that live facial recognition is not currently used in Northern Ireland. I was aware of what the hon. Member for West Dunbartonshire (Douglas McAllister) said when he intervened earlier, because Northen Ireland is in the same place on this. The FOI concluded that it is the intention of the PSNI to explore fully the potential of facial recognition technology, and that a working group was to be established in late 2022, in conjunction with PSNI stakeholders. The principle of why the technology is necessary is already in place, but we need to have the safeguards as well.
Last week I was in a Westminster Hall debate secured by the hon. Member for North Down (Alex Easton) on the importance of funding for local policing. There are clear examples across the United Kingdom that show that live facial recognition works and is extremely beneficial to the prevention of crime and for convictions. Perhaps, then, it is something that could be funded through the Barnett consequential. The Government will tell us that they have set funds aside, and we thank them for the extra money for Northern Ireland, but if it can help the police forces, that needs to be looked at.
Numerous concerns have been raised about the use of LFR by our police forces. Surveys have revealed that the British public are mainly concerned with policy infringements, surveillance, consent and the unethical use of facial recognition by the police. The hon. Member for Brent East said that clearly in her contribution, as have others. Police officers shared concerns that there could potentially be impacts on the legal and human rights of citizens.
I will always speak out on human rights abuses where they are highlighted and where infringements take place. It is good to see the Minister in her place; we all have an incredible respect for her and I look forward to her contribution. I seek to hear from her how human rights can be assured and carefully covered. The invasion of liberty and privacy are of major concern. If the technology is to be widely used across police forces, there must be assurances on public safety.
Concerns about false positives have been raised. I do not pretend to understand the technology, but others have explained that if it is turned down from 0.6 to 0.5, it offers a wider spectrum of people. That can cause such damage to people and their reputations, and reputation is everything for many people. Should this be trialled in the likes of Northern Ireland or Scotland, we must have assurances that the algorithms are correct and that they identify people correctly. I support the technology with that proviso.
Like many others, I have many concerns about live facial recognition technology, some of which have already been raised, but I will focus my remarks on the room for error and the potential impact that this technology will have on already dwindling public trust in police, particularly among black, Asian and ethnic minority citizens. I will raise points similar to those of my hon. Friends the Members for Liverpool Riverside (Kim Johnson) and Brent East (Dawn Butler).
Live facial recognition technology compares live CCTV images with those already on the police database and other images taken from open source, publicly available image sites. This is a deeply flawed plan that could result in serious mix-ups. A simple mislabelling on an image database could lead to the wrong person being stopped and a potentially traumatic experience with the police.
I can illustrate my point with a short anecdote; this happened to me a mere few months after I was elected to this House. My hon. Friend the Member for Battersea (Marsha De Cordova) was speaking in the Chamber. BBC Parliament miscaptioned her as my hon. Friend the Member for Brent East and, when they spotted this, both Members took to Twitter to point out the mistake. In their haste to cover the story, the Evening Standard incorrectly used a picture of me instead of my hon. Friend the Member for Battersea—I hope everybody is following this—and in its apology to all three of us, it suggested that Getty Images, where they had taken the image from, had labelled most of the pictures of me, since I had been elected, with the name of my hon. Friend the Member for Battersea. Since then, to avoid embarrassment, it seems that most publications now use pictures of me looking like a constipated walrus, but they have said that their reason for this is that they can be sure it is me and they want to avoid any further embarrassment.
Although problematic, that is a far more trivial example of what can happen when images are mislabelled, but if humans can make these errors, the technologies they create obviously can. If online sources are going to be used as part of the image database, it is almost inevitable that images will be mislabelled and that innocent people will be subject to needless run-ins with the police.
Questions around the numerical similarity score used to determine matches also ought to be raised. We already know that facial recognition data has racial bias: it is deeply flawed when attempting to identify people with darker skin tones, just as Getty Images is, and the Metropolitan police’s own testing of its facial recognition algorithm identified disproportionately higher inaccuracy rates when attempting to identify people of colour and women.
People of colour are already disproportionately stopped and searched at higher rates, and the use of potentially flawed technology will serve only to increase the rate at which ethnic minorities are stopped, searched and possibly even incorrectly detained, further dampening trust in the police among these communities. We know that that needs to be resolved. To any Member who thinks that I am exaggerating the potential for misidentification, I say this: in 2023, Big Brother Watch found that over 89% of UK police facial recognition alerts wrongly identified members of the public as people of interest. In that case, what benefits does this technology bring? It has been used in the borough of Lambeth, including in my own constituency, on a number of occasions, but as far as I am aware it has not produced a substantial number of results. Our constituents are effectively being placed under constant surveillance. The notion of their presumed innocence, which sits at the heart of our justice system, has been undermined, and this “cutting-edge” technology has not produced substantial results.
With some 6 million CCTV cameras in the UK, which all have the potential to be converted into facial recognition cameras, we are veering dangerously close to becoming a police state with levels of surveillance that would be deemed acceptable only in the most authoritarian of dictatorships. I believe that our liberty and our security can co-exist. It is not a matter of “those who have nothing to hide have nothing to fear”; it is a matter of the basic principles of freedom and privacy. Those basic principles begin to draw into question what such surveillance is really here for. Is it here to keep us safe or to monitor us 24/7?
Most Members would undoubtedly, I hope, protest at the idea of police randomly stopping members of the public to check their fingerprints or other DNA against databases just for a possible match. Why should we look at this intrusive automated biometric software any differently?
It is said that technology is a very useful servant but can be a very dangerous master. Many colleagues have already made a robust case for the use of this technology and undeniably it can be very useful. However, I am extremely concerned and believe that we must proceed with caution. In Leicester, some people already want to use the technology, but we must ensure that there is watertight legislation before we proceed any further.
Among my main concerns is the accuracy of the technology. We must ask whether it is fit for purpose. A spokesperson from StopWatch, a UK coalition of academics, lawyers and activists, has said that
“there is very little evidence on the efficacy of LFR deployments”.
In fact, in the first six months of this year, when this technology was deployed, StopWatch found that on average it stopped one person nearly every hour, or every 55 minutes, and that a person was arrested every two hours because of it. The data showed that, as the hon. Member for Clapham and Brixton Hill (Bell Ribeiro-Addy) said, over 80% of those arrests were unnecessary. The right hon. Member for Maldon said that the police have polite conversations with people, but polite conversations have a different meaning for different people.
Secondly, there is equality and non-discrimination. We already know that a black person is four times more likely to be stopped by this technology, as we are now. The technology has been shown to exacerbate any racial profiling. In fact, it has been demonstrated that it disproportionately misidentifies women, people of colour and even disabled people. That is a real concern.
Thirdly, as the majority of colleagues have already mentioned, the technology is an attack on our civil liberties. Earlier this year, the European Court of Human Rights ruled against Russia after claims that it had used LFR technology to locate and arrest a protester on the Moscow metro system. That is extremely frightening. Similarly, China has been accused of perfecting a version of facial technology that can single out and track Uyghurs—members of the repressed Muslim community in China.
We must acknowledge these concerns and ensure that, like the EU, we have in place stringent legislation, like the EU’s Artificial Intelligence Act 2024, before this technology becomes widely used and turns into our master.
I have worked on this issue for many years. In my previous job, I attended and observed the first deployments of live facial recognition by the Metropolitan police, which is many years ago now. Since then, the gap between its increasing use and the lack of a legislative basis has grown wider and wider. In that time, many thousands of people have had their personal data captured and used by the police when there was absolutely no reason for that. Many people have been misidentified, but the accuracy issue is not my main concern.
The unlegislated use of the technology is incredibly worrying. In my previous job on the London Assembly, I asked the Met and the Mayor of London many questions about that. I asked for watchlist transparency, but I did not get it. I heard the initial promises—“Oh, it will be very transparently used, we will communicate it, and no one will have to walk past it without knowing.” All those reassurances just faded away, because there is no real scrutiny or legislation. We need to debate the subject from first principles. As other Members have pointed out, we have had proper debates about identity cards and fingerprint and DNA data, but not about this extremely intrusive technology. It is more concerning than other technologies because it can be used on us without our knowledge. It really does engage our human rights in profound ways.
For all those reasons, the use of facial recognition by the police has been challenged by the Information Commissioner, the Surveillance Camera Commissioner, the Biometrics Commissioner, London Assembly members, of whom I was one, Senedd Members and Members of Parliament here. The only detailed scrutiny of the technology has resulted in calls for a halt to its use; I am thinking of the Science, Innovation and Technology Committee. The Justice and Home Affairs Committee has also called for primary legislation. That is the absolutely key question. The EU has had the debate and looked at the issue in detail, with the result that over there what is used so much by the UK police is restricted to only the most serious cases of genuine public safety. That absolutely needs to happen here.
The legislation needs to look not just at police use of the technology, but private use. I have seen its use by private companies in the privately owned public space in King’s Cross. Data from there has been shared with the police; the police initially denied knowing anything about it and then later apologised for that denial. If private companies are collecting data and sharing it with the police, that needs to be scrutinised. If private companies are using the technology, that needs to be legislated for as well.
I tabled a written question to the Minister about putting regulation and legislation behind the police use of live facial recognition. The answer stated that the technology is governed by data protection and equality and human rights legislation, and supplemented by specific police guidance. I do not believe that police guidance is sufficient, given the enormous risks to human rights. We need a debate on primary legislation. I hope that the Minister will announce that that process will start soon and that this unlawful grey area will not be invading our privacy for much longer. This issue is urgent.
I am grateful that concerns have been raised about how the technology we are discussing impacts the assumption of innocence—we should all be very careful about that—although I also appreciate the point that it does not impact innocence but provides the opportunity for a human to check. If done properly, that is no bad thing, but we are right to discuss the issue in serious terms in our legislature because there is a danger of an unofficial assumption of guilt. Let us take the example of local shopping centres, which we heard about earlier. If an issue has not been escalated to the police or courts, but some local security officers have seen the same images on cameras and that information has gone round by radio, a gentleman or a lady out with their children doing the weekly shop may suddenly not be able to get in and do what they need to do. That is the kind of pervasive and damaging thing that could easily slip under the radar; we should all be mindful of that.
I want to touch briefly on transparency. This is clearly a developing technology and we would be wrong not to look at its benefits, but we must be mindful of the harm it could do along the way. If people find that they are getting an unfair crack of the whip—that is probably an inappropriate term—and are suffering as a result of this technology, we need to nip that in the bud, and be very direct and open about the failures so that we can make adjustments.
To go back to transparency, we have to be open and frank about any issues with how the technology is being implemented, so that we can fix them. I agree that there absolutely could be issues, and we definitely want to be on the right path.
The hon. Member for Brent East (Dawn Butler) used the excellent analogy of a night out. I completely agree; I was thinking, “Yeah, I’m up for it, but let’s just make sure we can all get home safe”, but the more we discuss the issue, the more I think the appropriate camp to be in is, “I could be tempted out, but let’s make sure we like the destination.” I will leave it there. I thank hon. Members for their time.
I have researched this subject and listened to hon. Members’ contributions, and it has been frankly shocking to learn that LFR has been in use since 2017 without any specific legislation in place to control its use and protect our civil liberties. That is seven years too many without legislation. Although I agree that the use of real-time facial recognition in the United Kingdom promises enhanced security and efficiency, it also raises significant legal and moral concerns, and there are severe adverse consequences for our society.
As a former software test manager, I am extremely concerned that private companies that profit from their technology are allowed to self-regulate and to confirm the efficacy of the products that they sell, and that the police are guided by those companies in how to use the tools and rely on the companies’ reports of their efficacy to take legal action against innocent civilians.
The technology operates by capturing and analysing highly sensitive and personal biometric data. As has been mentioned, the legal framework for its use is complex and at times insufficient. The Data Protection Act 2018 and the General Data Protection Regulation provide some safeguards, requiring data processing to be fair, necessary and proportionate. However, the lack of specific legislation for facial recognition technology leaves huge room for misuse and overreach.
The deployment of this technology without explicit consent undermines several of our fundamental rights, some of which have been mentioned. The first is the right to privacy: constant surveillance and the collection of biometric data without explicit consent infringe an individual’s privacy rights. This is particularly concerning when the technology is used in public spaces without people’s knowledge. The second right is the right to freedom of peaceful assembly and expression. The use of facial recognition can deter individuals from participating in protests or public gatherings due to the fear of being monitored or identified. This undermines the fundamental right to assemble and express opinions freely.
The third right is the right to non-discrimination. As has been mentioned, facial recognition systems have been shown to have higher error rates for people of colour, women and younger individuals. This bias can lead to disproportionate targeting and wrongful arrests, exacerbating existing inequalities and discrimination. The final right is the right to data protection. The collection, storage and processing of biometric data must comply with data protection laws. Inadequate safeguards can lead to unauthorised access and misuse of personal data.
My hon. Friend the Member for Leicester South (Shockat Adam) cited examples of how this technology is used in Russia and China, and we know that it is used extensively in Israel as part of its apartheid regime and occupation of the Palestinian people. Violations highlight the need for strict regulation and oversight to ensure that the deployment of facial recognition technology does not infringe fundamental human rights. The technology subjects individuals to constant surveillance, often without their knowledge, eroding trust in public institutions. The ethical principle of autonomy is compromised when people are unaware that their biometric data is being collected and analysed.
Let me cite some examples of the technology’s inefficacy and unreliability. In 2020, the Court of Appeal found that South Wales police’s use of facial recognition technology was unlawful, and that the force had breached privacy rights and failed to adequately assess the risks to individual freedoms. The technology’s accuracy is not infallible: misidentifications can lead to miscarriages of justice, where innocent individuals are wrongly accused or detained.
The disproportionate impact of FR technology on black people and people of colour is particularly concerning. Research has consistently shown that these systems are more likely to misidentify individuals from those groups. For example, a National Institute of Standards and Technology study—I do not know how old it is—found that FR algorithms were up to 100 times more likely to misidentify black and Asian faces than white faces. This disparity not only undermines the technology’s reliability, but perpetuates systemic racism. In practice, this means that black people and people of colour are more likely to be subjected to unwanted surveillance and scrutiny, which can lead to a range of negative outcomes.
There are other examples of miscarriages of justice and misuse. In one instance, the Metropolitan police used FR technology at the Notting Hill carnival, leading to the wrongful identification and harassment of innocent individuals. These and the other examples cited by hon. Members underscore the potential for significant harm when this technology is deployed without adequate safeguards.
In conclusion, although facial recognition technology offers potential benefits, its deployment must be carefully regulated to prevent misuse and protect individual rights. The legal framework needs to be strengthened to ensure that the use of technology is transparent, accountable and subject to rigorous oversight. We must also address the inherent bias in these systems to prevent further entrenchment of racial inequalities. As we navigate the complexities of integrating new technologies into our society, let us prioritise the protection of our fundamental rights and ensure that advancements serve to enhance rather than undermine our collective wellbeing.
What concerns me is the idea of allowing the state, in essence, to deploy this kind of technology in high streets, for example. The hon. Member for Brent East (Dawn Butler) has already raised the issue of the disproportionate rate of stop and search—by multiple times; I think the rate was nine times higher for black males. What impact will facial recognition live transmission data have in the city of Birmingham? It is going to have an enormous impact. Members have raised the difficulties with the percentage error of recognitions, and the distrust that we have in Birmingham is a challenge already, particularly with young men and the police. What will this technology achieve? Will young men start wearing more face coverings in city centres? How will this technology be used, even if it is legislated for properly? For example, will the police have to notify the public, “We are using this facial recognition technology in the Bullring today between the hours of 10 am and 10 pm?” It does not seem to serve any real purpose.
We have a very effective police force in the west midlands, and it uses CCTV, which we have all over. If hon. Members go to any street in Birmingham, they will find tens or hundreds of houses with CCTV, and the police have used that to great effect; after a crime is committed, they track back and they prosecute. We have had so many successful prosecutions in very serious crimes, such as murder and violent crime, but the deployment of this technology will create enormous problems and divisions. As I said, there are already problems with how minority communities feel when they are stopped and searched. I think the right hon. Member for Maldon said that in the trial about 10 people were stopped, with one to two—as little as 10%—being identified. As the technology develops further, that percentage may increase, but at present I do not see how it will assist at all. Criminals know very well how to avoid detection, and face coverings will become the norm. Other than surveillance, this technology achieves very little. I do not see how it will assist in detection.
The hon. Member for Brent East drew some simple parallels. What would the public think about being stopped on a busy high street and asked to come to a police van to give their fingerprints and DNA? They would be outraged, and rightly so. It would almost legitimise police officers approaching people, in particular young men. We know that not just black people, but people of colour, women and children will be subject to the technology, and we know that there are errors. The right to privacy and the freedoms that we have are far greater than this technology, and I do not see how it will assist in deterrence, because people will simply use face coverings and all sorts of other things.
We have discussed whether live facial recognition technology is a legitimate tool and, if so, under what circumstances and controls it should be used. It is clear from the debate that there are many doubts, and we should probably be thinking about halting the use of the technology until we have cleared them up.
I will start with the concern about discrimination, which was articulated well by the hon. Member for Brent East (Dawn Butler). It is clear that black people and people from other communities are likely to be disproportionately misidentified by this technology.
The hon. Member for Leicester South (Shockat Adam) made a comment about how polite conversations do not always register as polite conversations. That is because of the persistence of those conversations over time. A repeated polite conversation starts to become an aggressive conversation to the person on the receiving end, if it is that persistent. There was also discussion about the findings of the national physical laboratory, but it is clear that those findings are disputed—[Interruption.] Well, it is clear that they are disputed; they have been disputed in the Chamber today. Until we get to the bottom of that, we need to think carefully about the controls that we have in relation to discrimination.
I want to talk about the general principle of privacy. As a liberal, I feel a general depression about how we have come to devalue privacy in society, and how we trade it away far too readily for other societal aims. We often hear the claim, “If you’re not doing anything wrong then there’s nothing to worry about,” as if the only value of privacy were to hide things that someone might be doing wrong. That is not the case. Privacy delivers so much more than that. It delivers personal wellbeing and gives people control over their own data. It allows us to have freedom of association and dignity. We need to think very carefully before we so readily trade away the principle of privacy in pursuit of other goals in society.
The opportunity for slippage has been discussed at length. One would think that such technology would come with strict controls, but it is clear that at the moment we have the opposite; in fact, Big Brother Watch has described it as a “legal vacuum”. The hon. Member for Brighton Pavilion (Siân Berry) talked about the creeping expansion of its use in London. I have seen that myself; what started off being limited to large-scale events, such as football matches, has turned into routine trials on high streets, such as mine in Sutton.
We have also seen expansion in the photos that are used. The technology started off using only photographs of people known to the police, for good reason, but it has been expanded to potentially including everyone who has a passport or driving licence photo. What started being strictly about warrant breakers and sex offenders could expand to be about pretty much anything the Government of the day want. If we think about the clampdown on protest under the previous Government, that potentially has a chilling impact on the right to freedom of association.
With all of those doubts, it is clear that we need proper parliamentary consideration of the issue. The Lib Dems ask the Minister to immediately halt the roll-out of live facial recognition technology until we get it right. It should be down to this place to determine the correct controls and whether there is a legitimate use of the technology at all, given all the concerns about discrimination and privacy. Privacy is a fundamental civil liberty. We have undervalued it far too much in recent times. This is an opportunity to protect it, and we should take it.
I think this is the first time that I have appeared opposite the new Minister for Policing, Fire and Crime Prevention—the job that I was doing until a few months ago—so let me congratulate her on her appointment. Although I will of course hold the Government to account, I will do everything I can to constructively support her in making a great success of the job, and I really do wish her well in the role.
I want to start by reminding colleagues of the way that live facial recognition works. It is different from retrospective facial recognition, which we have not debated today and, in the interests of time, I do not propose to go into. As some Members have already said, live facial recognition starts with a watchlist of people who are wanted by the police. It is not the case that anyone can get on that watchlist, which generally comprises people who are wanted for criminal offences—often very serious offences—people who have failed to attend court, and people who are registered sex offenders, where the police want to check that they are complying with their conditions. As people walk down a high street, they are scanned, typically by a CCTV camera on a mobile van, and then compared to the watchlist. The vast majority of people are not on the watchlist, as we would expect, and their image is immediately and automatically deleted. Where a person is on the watchlist, the police will stop them and ask if they have any form of identification.
To be very clear, no one gets convicted on the basis of that facial recognition match, so it is not overturning the presumption of innocence, and if it turns out that the person stopped is not the person on the watchlist, obviously they can continue on their way. However, if they are the person on the watchlist, a normal criminal investigation will follow, with the normal standards of evidence.
Let me come to the question about racial disparity. When this technology was first introduced, about seven years ago, there were reports—accurate reports—that there was racial bias in the way that the algorithm operated. The algorithm has been developed a great deal since those days, and it has been tested definitively by the national physical laboratory, the nation’s premier testing laboratory. NPL testing is the gold standard of testing and this technology has been tested relatively recently. For the benefit of Members, I will read out what the results of that testing were:
“The NPL study found that, when used at the settings maintained by the Met”—
that is the 0.6 setting that the hon. Member for Brent East (Dawn Butler) referred to earlier—
“there was no statistically significant difference in the facial recognition technology’s accuracy across”
different demographic groups. In other words, the technology as it is being used today—not five years ago, when there were issues—has been certified by the NPL and it has been found that there is not any racial bias at the settings used.
About 25% to 30% of regular physical stops and searches, where a police officer stops someone and searches them for drugs or a knife or something, are successful. About 70% are unsuccessful, while the equivalent figure for live facial recognition is 0.02%. That means that this technology is 4,500 times less likely to result in someone being inappropriately stopped than a regular stop and search. It therefore hugely—by three orders of magnitude—reduces the likelihood of someone being improperly stopped and searched.
I turn to the use of the technology on the ground. I asked for it to be trialled in the centre of Croydon, which is the borough I represent in Parliament. Over the past nine months or so, it has been deployed on a relatively regular basis: about once a week. I believe that the Minister was supposed to go down this morning to have a look; I certainly encourage her to go again as soon as she can. By the way, the hon. Member for Birmingham Perry Barr (Ayoub Khan) asked whether people know when the technology is being used. The answer is yes: one of the guidelines is that public signage must be displayed telling the public that the technology is in use.
Over that period in Croydon, there have been approximately 200 arrests of people who would not otherwise have been arrested, including for all kinds of offences such as class A drugs supply, grievous bodily harm, fraud and domestic burglary. It has also included a man who had been wanted for two rapes dating back to 2017. That wanted rapist would be free to this day if not for this technology. Just a couple of weeks ago, a man was stopped and subsequently arrested in relation to a rape allegation from June this year. There are people who are alleged to have committed rape who would not have been stopped—who would still be walking free—if not for this technology. It is only the fact that they walked past a camera outside East Croydon station or somewhere that has meant they were stopped by the police. They will now have a normal trial with the normal standards of evidence, but they would not have been caught in the first place if not for this technology.
I have done quite a lot of public meetings on this. I explain, “These are the people who get caught, and the price the public pay is that you might get scanned when you walk down Croydon High Street, but if you are innocent your picture is immediately deleted.” By and large, the overwhelming majority of the people in Croydon think that a reasonable trade-off.
There should be protections, of course. Several hon. Members, including my right hon. Friend the Member for Maldon, have rightly said that there should be guidelines, rules and procedures. However, it is not true that there is a complete vacuum as far as rules and regulations are concerned. The Bridges case at the Court of Appeal in 2020 looked at how South Wales police were using the technology between 2017 and 2020. It found that some of the ways they were using the technology were not appropriate because they broke rules on things like data protection privacy. It set out in case law the guidelines that have to be adhered to for the technology to be lawful—things like public signage, the rate of accuracy and having no racial bias.
Secondly—I do hope I am not taking the Minister’s entire speech—there are guidelines for police. The College of Policing has national authorised professional practice guidelines that the police are supposed to stick to. There is a debate to be had about whether, for the sake of clarity and democratic accountability, we in Parliament should set something out more formal; my right hon. Friend the Member for Maldon made that point. I think there would be some merit in clarifying at a national level where the guidelines sit, but I would not go as far as Europe. If we had done so, those rapists would not have been arrested. I would also be careful to ensure that any legislation is flexible enough to accommodate changing technology. Primary legislation may not be the right vehicle: a regulation-making power might be a more sensible approach, so that things can be kept up to date from time to time.
While we consider that, I strongly urge the Minister not to halt the use of the technology. As we speak, it is arresting criminals in Croydon and elsewhere who would not otherwise be caught. I urge her to continue supporting the police to roll it out. I think some money was allocated in the Budget for the current financial year, to continue developing the technology. I would welcome an update from the Minister on whether that money is still being spent in the current financial year. I do hope it has not somehow been snaffled by the Treasury in a misguided cost-saving effort—
I congratulate the right hon. Member for Croydon South (Chris Philp) on his new role. When he held the role that I now hold, he was very passionate about this subject. That passion is demonstrated today by the number of interventions he has made and by his contribution in defence of the previous Government’s approach to this particular policy. Now that we have seen the shadow Home Secretary in a Westminster Hall debate on this issue, I very much hope that we might see him here again when we debate the many other policing issues that we have to deal with, including police reform and police accountability—the list goes on.
This has been a very good debate. We have ranged from discussing the Jason Bourne films to a night out with my hon. Friend the Member for Brent East (Dawn Butler). We have also had excellent contributions from the hon. Member for Strangford (Jim Shannon), my hon. Friend the Member for Clapham and Brixton Hill (Bell Ribeiro-Addy), the hon. Members for Leicester South (Shockat Adam), for Brighton Pavilion (Siân Berry), for South Basildon and East Thurrock (James McMurdock), for Dewsbury and Batley (Iqbal Mohamed) and for Birmingham Perry Barr (Ayoub Khan), and the Liberal Democrat Front-Bench spokesperson, the hon. Member for Carshalton and Wallington (Bobby Dean).
I will deal with the complex issues that Members have set out so eloquently. At the heart of the issue is the fact that we are dealing with a powerful technology that has the potential to be transformational for policing. However, some have very legitimate concerns about it, including misidentification, misuse and the effect on human rights and individual privacy. I agree wholeheartedly that we need a proper, informed debate on the subject, both in this House and with the public, and I am pleased that we have had the opportunity to start that today.
Let me quickly run through the current use and benefits of live facial recognition, which, as we have discussed, allows the police to spot people in crowds. It uses live video footage of crowds passing a camera and compares their images to a specific watchlist of people wanted by the police. As well as Essex police, who we have heard about, the Metropolitan police and South Wales police have been using this technology for a number of years. In fact, as the shadow Home Secretary said, I was due to go and see one of the deployments this morning, but then this debate was scheduled, so I am going to see it for myself this afternoon instead.
I am told by the Metropolitan police that between January and November this year they made over 460 arrests as a result of live facial recognition deployments, including for offences such as rape, domestic abuse, knife crime and violent robbery. In addition, over 45 registered sex offenders have been arrested for breaching their conditions. South Wales police tell me that between January and November, they deployed live facial recognition locally on 20 occasions, resulting in 12 arrests. They also located a high-risk missing young girl, who they were able to safeguard from child sexual exploitation and criminal exploitation. Essex police, as the right hon. Member for Maldon has attested, have also had considerable success in their use of this technology.
The potential of live facial recognition to contribute to our safer streets mission is clear. It could make our streets safer for us all, particularly for women and girls, by helping the police to identify wanted people quickly and accurately. It could also save precious police time. Rapid advances in the technology and improvements in the accuracy of algorithms increase that potential.
Let me consider the concerns that have been raised. I was pleased to hear that the right hon. Member for Maldon was impressed by the strict limit that Essex police have put on their use of live facial recognition. That includes use of the narrowly drawn watchlist and the immediate deletion of images. However, I note his worries about the lack of a specific legal framework for the technology’s use. It is therefore important to be clear that facial recognition is covered by data protection, equality and human rights law as well as common law powers and detailed guidance from the College of Policing. However, the right hon. Member is right that no one specific law gives the police the power to use live facial recognition.
The Ada Lovelace Institute, an independent research institution with a mission to ensure that data and AI work for people and society, has written to the Home Secretary to express similar concerns to those of the right hon. Member. It believes that the only way to scale up those technologies safely and successfully is through the introduction of a statutory regulatory framework. I have spoken to senior police leaders about the matter, and some believe that the lack of a specific legal framework inhibits their use of the technology and dampens willingness to innovate.
With legal challenges highly likely, it is not surprising that some police forces are reluctant to use the technology. However, others in policing are keen to emphasise the safeguards that are already in place. For example, they assure me that the police do not keep the biometric data of people filmed during live facial recognition deployments, that watchlists are bespoke and that the police deploy the technology only when there is an intelligence case for doing that. I have also been assured that there will always be a human being in the loop to decide whether to apprehend someone. That would never be done solely on the basis of a match made by a computer.
Privacy campaign groups have a long-standing interest in the subject. I am aware of their concerns, as well as previous and ongoing legal actions relating to police use of live facial recognition technology. Potential bias in the algorithms used for live facial recognition systems is another frequently raised concern. Questions have been asked today about that very point and whether live facial recognition discriminates against people on the grounds of gender or race. I am also aware that 65 Members of Parliament and peers signed an open letter last year that called for a ban on live facial recognition, and that in January the House of Lords Justice and Home Affairs Committee sent the then Home Secretary a report raising concerns and making recommendations about live facial recognition.
I remind Members that the Government have been in post for five months. Let us put that in the context of the previous 14 years of Conservative Administrations. The Government want to take time to listen and to think carefully about the concerns that have been raised and about how we can best enable the police to use live facial recognition in a way that secures and maintains public confidence.
As we have heard today, facial recognition technology is a powerful tool. In considering its current and future use, we must balance privacy concerns with the expectation that we place on the police to keep our streets safe. We particularly need to consider how much support the police may require from Government and Parliament to set and manage the rules for using technologies such as facial recognition. We must think about how we protect the public from potential misuse of those technologies, and we need to consider how the application of the rules and regulations is scrutinised.
I am therefore committed to a programme of engagement in the coming months to inform that thinking. Building on initial conversations with police, I will hold a series of roundtables, for example, with regulators and civil society groups before the end of the year. I look forward to hearing at first hand from a broad range of parties on the subject.
I am running out of time. I want to say much more on this issue, and I want to confirm that money is being spent this year on the roll-out of the live facial recognition vans that are being equipped to carry out this work. There is a full evaluation of that work going on. I very much look forward to the House having further opportunities to debate the issue in the coming weeks and months.
Motion lapsed (Standing Order No. 10(6)).
Contains Parliamentary information licensed under the Open Parliament Licence v3.0.