All right. So, hello, everyone. I am Jaye. I am the Chief Technology Officer of a tech-for-good company called Vouchsafe, and we're kind of experimenting with what we say that we do in settings like this, but this week I'm trying out the phrase trust infrastructure. This is us. That's my co-founder, Chloe, to my left there. Trust infrastructure is, again, it's just a term that I made up.
So, what that means to us is trying to build more flexible, more inclusive identity tools that solve for those underrepresented use cases that are going unnoticed by the industry at large, and there are quite a lot of them, and also trying to get away from that point in time, send a user to us, we verify them, then we send them back to you kind of journey that I think everyone is hooked on at the moment, and try and get more towards solving a whole problem for a business, including, but not limited to, an ongoing picture of risk. So, my personal career before Vouchsafe was pretty varied.
I had a really long stint in government working in tax and then universal credits, which are the two biggest single IT systems that the UK government runs, and I was the original designer of the UK's coronavirus contact tracing app in 2020. I'm sure we remember those. I've also spent time at some big charities like Samaritans, the British Film Institute, I've worked as an engineer, a designer at a bunch of startups, and originally I was a journalist.
So, I like to think that I've seen a fair sampling of the industry, and one of the things that I saw again and again and again in all of those roles is people struggling to prove things about themselves. Working at all these places, I actually built many versions of the same tool, which is basically something to collect paper evidence and some like janky dashboard interface that some caseworker or member of staff would just look at with their human eyes and just hit approve.
So, I'm going to make this talk comparatively light on stats, not least because if you're in this room you probably already know the most important ones, but it will be quite heavy on case studies, specifically case studies from our three years in the field, so you probably won't have heard any of them before. Not just researching the problem of ID poverty, but like trying to engineer solutions to it that you can apply in isolation and have a positive social impact right then and there.
So, let's get those stats out of the way up front. When I talk about ID poverty, this is not the only definition, but when I talk about it, I'm talking about the inability to prove who you are no matter the cause, and that's deliberately more of a customer experience kind of framing than an international development humanitarian kind of framing, just because of who our customers are, right, and the kind of stuff they care about.
That could be proving who you are for something important, like we've heard about in this track, like a bank account or a government service, or it could be for something much more trivial. The World Bank number, which is on these handouts that are dotted around the room, 1.1 billion globally in this situation, and this of course is specifically the number that the World Bank estimates lack official government ID.
So, for our purposes, this is probably an underestimate. In the UK, where I'm focusing on for this talk, the number is 11 million. Specifically, this is the number of adults who don't have either a passport or a driving license.
The UK, if you're not familiar, doesn't have a national ID scheme, so those two documents are pretty much it. This number wasn't particularly well understood until our last general election last year, where we introduced voter ID, and suddenly it became incredibly democratically important to know what this number was. And the shape of the problem is changing, it doesn't exist in a vacuum, it's a really beautiful example of the poverty premium, that is to say, demanding more of people who have less and providing them a worse service.
It's also naturally related to financial exclusion as well, and it's being made worse by the cost of living crisis, which has hit the UK much harder than it's hit mainland Europe in most cases. Fewer people can simply afford to travel internationally now, and fewer people are learning to drive.
In fact, the average age where you learn to drive in the UK now is 26, and you'll have done most of this stuff by the time you turn 26 that is asking you for a driving licence at least once. So let me give you these case studies now, let's take a look at how this is manifesting in the real world, shall we? So officially, you're not supposed to need much to sign up for a family doctor in the UK.
Our National Health Service, quite rightly, takes the view that the more barriers they put in place to early care, the more it ends up costing them in the long run, because those people will inevitably hit a crisis that you weren't able to prevent. So you're certainly not supposed to need photo ID, for example, but one of the very first pieces of guerrilla research Vouchsafe did in the early days was to call round every GP surgery in the local area and just say, hey, I don't have this thing, what can you do for me? And every single one of them was like, oh no, you need that.
Like, I'm not sure what we can do if you don't have that. They just didn't have a process for it. It is a similar case for all sorts of government services. My co-founder, Chloe, led user research on universal credit for three years. UC is the chief government all-in-one benefit in the UK. It replaced JobSeeker's Allowance and all sorts of other things. And it is notorious for the amount of paperwork you need to provide to evidence a claim.
And Chloe describes seeing people bringing in literal binders worth of paperwork to meet with their work coach in order to hopefully get their claim approved. Often this is driven by people just not having that clean, high-quality evidence that they're assumed to have, so they have to resort to this old-fashioned certain number of things from list A, certain number of things from list B kind of process, all within the last six months, ideally.
One of the things that Chloe did at universal credit was helping people who didn't have a fixed address claim their benefit, because many of them didn't have bank accounts, because it's really hard to open a bank account without a fixed address. One of our customers now, one of our partners, is a payment processor called Allpay. And one of the things they do a lot of is paying out money from local authorities, dispersing money for various purposes.
And Allpay noticed, to their credit, that a lot of their customers didn't have a way to receive that money, often because they didn't have the prerequisites to pass the bank's KYC process. Banks that have been around for a long time have this problem particularly badly as well. One interesting thing we've learned about is that many of them have large numbers of older customers on their books who were grandfathered in, basically, from times where the KYC was basically like walking into a branch and like, you know the guy, so you know, what are you going to do?
And now, of course, these banks are all trying to modernize and take these people through their shiny digital KYC and get them on their apps, but they're finding they just don't have the documents to do it. So what do you do? Do you shut down Betty's bank account?
Ideally, they'd prefer not to do that, but what do you do instead? Often, it must be said, these customers are older women, which poses a problem for lasting power of attorney, for instance, which we've learned from some research with the Ministry of Justice. Power of attorney is obviously such a big decision in a person's life that you have to be super sure you're dealing with the right people, but many of these older women come from the era where you couldn't open a bank account as a woman until like the mid-70s.
Interestingly, though, something they do often have are very strong social connections, which we will come back to. Local credit unions also see a very extreme version of this problem.
Again, credit unions, I find, like, sometimes I walk into a room and say credit unions and everyone knows what they are and sometimes no one does, so who knows what a credit union is? Yeah, okay, just imagine like a really tiny bank, basically, like a really tiny bank, like an old-fashioned bank where it was all done in branch and it's all paperwork in the back. That's basically what a credit union is in the UK. We've toured the country over the past few months talking to different credit unions.
This is Advance in Birmingham on the left there, which is about 10 minutes from the street I grew up on, and then this is Rightway on the right, which is in Paisley, just outside Glasgow. Credit unions are a really good ethical alternative to those predatory high APR lenders that people would otherwise have to go to. They give small loans to people who would often not be able to get mainstream credit, and that segment naturally has a much greater degree of ID poverty. That segment is also growing as the cost of living pushes more people toward needing credit for everyday essentials.
So, where before your family finances might have been able to weather, paying for a new boiler, say. The UK is horrendous for the state of our central heating and HVAC and all of that. These days you'll probably need the help of your local credit union, and the state of the art for technology in the credit union leaves a lot to be desired, which doesn't help. One place with a workable solution to this is CitizenCard. This is a proof-of-age card that young people use to go to the cinema and buy alcohol and all sorts of things like that. There are about 400,000 of them around.
It's issued by a company also called CitizenCard, and because it's often your first form of photo ID as a young person, you probably won't have another ID that you can use to get it, and you certainly won't have, say, a credit file. So, what do you do?
Well, CitizenCard has partnerships with schools, universities, and even prisons where members of staff vouch for these cohorts of people coming through, and that by itself is enough to break the catch-22 and get things moving, but it is a hand-cranked manual process where these referees are literally called up on the phone. A landline phone, it has to be, otherwise it doesn't prove that they work at the place that they say that they work.
So, the common themes to all of this. Firstly, there's a lot of confusion about the rules in practice, to the point where it doesn't really even matter what the rules are, because there's such a degree of confusion that everyone is just applying them based on what they reckon anyway, and it is very clear to us that this is only the tip of the iceberg, right? These digital pathways, these digital solutions that we talk about a lot at conferences like this, they're just the start.
It is still far too easy in the real world to fall back to these manual paper-based processes where the decision still depends on just what someone reckons, and as the identity ecosystem and the standards that back it become more interconnected, that is just not going to be practical for much longer. So, we do need a better solution to this. Who's seen this poster before?
Yeah, give me a hand if you've seen this before. Interesting.
Okay, this is a poster from Microsoft about accessibility. It's making the point that when we think about access needs, we normally think about the permanent ones on the left, right?
One arm, blind, deaf, non-verbal, but there are also temporary and situational disabilities to consider. You might be deaf, or you might just have an ear infection, or you might just be in a loud place. These are all disabilities, and ID poverty is a similar kind of thing. There's a spectrum of need here, and we are very used to thinking of people at the extreme left end of this spectrum, but then there's also people who say, just don't have it on them right now, on the right end. I was in the latter group quite recently.
I changed my name, and I had to send every document I had off to the passport office in a little envelope, and during that two, three-month window, I was just stuck every time something asked me to prove who I was, and I had to go through a fundraising process for the company during this period, and every single one of those investors is trying to do the video selfie ID scan on you, and it was just impossible. So, it is really, really important to avoid thinking of this as just something that affects poor people, or just something that affects migrants.
That said, it is a problem that overwhelmingly affects young people, older people, women, anyone who's not a middle-aged, middle-class guy, basically, and we can actually start to see these patterns reflected in our service's own analytics, so we've got data to back this up. It's not just anecdotal anymore, and of course, many of these people are going to be using AI agents over the next few years, expecting them to be able to act on their behalf, making use of their sensitive data, and today's IDV is just not going to be good enough to cope with this, as we've heard from other talks today.
I think this is a particularly good reminder that it's not that the people I'm talking about can't access any services at all. That would be too simple. There is a small minority who truly can't, but for the majority who can, it is still an expensive, risky, frustrating hassle for both the service provider and the consumer. When you're on the business end, it can sometimes be hard to notice that you have this problem, because it is what I like to think of as an outside context problem.
It necessarily sits outside the systems you've created and your ways of monitoring them, but a few common symptoms are a low acceptance rate, or at least lower than you'd ideally like, a high retry rate, especially on steps like video selfies, and a high rate of manual efforts to get people onboarded, which tends to mean, just part and parcel with that, those people are going to get a second class service, compared to the people who can self-serve.
Barring some exceptions, like Citizen Card, who do things differently, the best verification rates we've seen in the real world, all in, are about 80% right now, and this is like, you know, government services that are really investing a heck of a lot of time and effort to try and get this as good as possible. For businesses that specialize in this segment, 60% is probably more typical, and we've seen like, you know, some of those credit unions I showed earlier, their rates are hovering at about 40% or even lower.
More than half of their customers are in this situation, simply just not having the correct things to prove who they are. And I'll reiterate Elizabeth's point, digital ID alone will not fix this problem. You may have seen that the UK government is now looking at digitizing identity documents, we're going to have mobile driving licenses very soon, but they are, crucially, they are not giving everyone a driving license or a passport. In order to verify a credential, the credential does need to exist first, right?
This is not a universal identity mandate or anything like that, and I suspect that that will continue to be as politically untenable in the UK as when Tony Blair tried it 20 years ago now. Which brings us back to vouching. One of the things that Vouchsafe does really well is what we've taken to calling biometric vouching. It is a modernized automatic process, automatic version of that same process citizen card have been using so effectively.
We back it up with biometrics and like modern anti-fraud tech so anyone can use it, and it meets the same risk profile and confidence threshold as the more widespread methods that you might be familiar with. We have successfully had this assured by the UK government now, so we're very excited, and we have found it to be an exceptional fallback solution to the problem that we've been talking about. Because even in the segments where ID poverty is worst, there are often still really good long-term high-trust relationships where one person can be a referee for another.
Like the non-profit you were talking about, I think, would be a really good example of this. This is a photo from three years ago. I'm holding the very first prototype for what became Vouchsafe in my hands there, and since then the need has only become more urgent, and I'm just really pleased that we've been able to not just study it but make some tangible progress on it together.
All right, so citations. A lot of this is internal research that we've done over the past three years, specifically research led by Chloe, my co-founder, who's on the bike behind me there, and some of it comes from a group called Project Nemo. They're really great. They're a financial inclusion advocacy group who came basically out of nowhere about six months ago, but they are making real waves in the UK right now.
So if you would like to learn more about this research, or if you have a service that doesn't work particularly well for people who aren't like 30 to 50 year old white guys with MacBooks, and you'd like to make it work better, then come talk to me. Questions? So I'll ask one. You're building a solution that will help for people that otherwise may not have a path. How do people get connected with you so that they know that's available to them at all?
Well, it came through the Scottish Government initially. So the Scottish Government, to their credit, they're a little bit further ahead than the UK Government are. They're a bit more Scandinavian in their outlook on this stuff, and they were seeing the problem that I described. They have about like an 80% success rate right now, so they're already doing pretty well, and they were like, we really want a way to get this last 20%, and we can't get it by scanning passports and driving licenses. We certainly can't get it by checking credit files. What else can we do?
Well, vouching. So with their cover and their political connections, we were able to come up with a way of doing it that's secure enough that they'd accept it, and because government in the UK is so obsessed with reusable identity now, they're trying to guard the entire ecosystem rather than just their own business purposes, which means that it's a much higher standard than if we started with like a bank, say, and now it works for everyone from God down. Interesting.
So I know that in the Commonwealth of Virginia, before 9-11, you could get a driver's license by having someone in the Commonwealth vouch for you. That was no longer true once we passed the Real ID Act. So it's interesting to hear that coming back.
I mean, it was always a practical methodology, if not abused. Yeah, I mean, we've been, I mean, we didn't invent this idea, right?
Like, this idea is as old as dirt. The thing that we've tried to do to it is to modernize it and make it a thing that you can do automatically and make it fit modern anti-fraud standards. But it is surprising how many like little examples come out of the woodwork when you look into this.
Like, I think, you know, voter ID in Saskatchewan and things like that. It's everywhere, but it's done a little bit differently in every place. All right.
Well, thank you very much. Very interesting.
