Ethics & Innovation: Can AI Close the Justice Gap?

Susannah de Jager: Welcome to Ethics and Innovation by Oxford Plus, a special miniseries hosted by me, Susannah de Jager and sponsored by Equinox, Equitable Innovation Oxford.

Today on Oxford Plus, I'm speaking to Professor Philippa Webb. Philippa is a professor of Public International Law at Oxford's Blavatnik School of Government, a barrister at Twenty Essex, and one of the world's leading scholars and practitioners in international law and human rights. Her work spans the International Court of Justice, the International Criminal Court, the European Court of Human Rights, and some of the most significant international legal questions of our time.

More recently, alongside Amal Clooney, Philippa co-founded the Oxford Institute of Technology and Justice, a partnership between the University of Oxford and the Clooney Foundation for Justice that seeks to harness artificial intelligence and new technologies to improve access to justice, strengthen accountability, and advance international law. At a time when technology is reshaping everything from our democracies to our legal systems. Phillipa's work asks a vital question, can these technologies help build more just and flourishing societies?

Today we'll discuss justice in the digital age, the opportunities and risks presented by AI and how technology might strengthen, not weaken the institutions and values upon which open societies depend.

Philippa, thank you so much for joining today.

Philippa Webb: It's a pleasure.

Susannah de Jager: AI is here and everything about it is moving at pace. Justice has been a little bit of an anomaly actually for adopting it in this country at least. I can't wait to hear from you of some of the things you're doing with the newly formed Oxford Institute of Technology and Justice. But just to contextualise for those listening, give us a little bit of your background and what brought you up to this point in your career.

Philippa Webb: Thank you. So I am a public international lawyer, which means that I'm concerned with relations between states. But as part of that, I'm also concerned with how states relate to their citizens, which brings in human rights. And I've had a career that sort of combined academia and practise. Practise both at the United Nations and the International Court of Justice and the International Criminal Court and then more recently at the Bar of England and Wales.

As part of my work, both in practise and academia, I've been focused on big fundamental rights, the right to a first trial, free speech, freedom from slavery. When working on these topics, the issue of technology was like the elephant in the room. How can you talk about fair trial without thinking about how AI may be transforming that for judges, for defendants, for prosecutors? And how can you talk about free speech without thinking about how the tech companies are approaching their obligations or pushing back on their obligations? So the time came a couple of years ago when together with Amal Clooney, who I had worked on the fair trial and the free speech books with, we said it's time to tackle the role of technology in this.

And we didn't think that looked like another book. We though it was something more practical, something more interdisciplinary and we came up with the idea of the Oxford Institute.

Susannah de Jager: So then go on. Tell us a little bit about the Oxford Institute, because it's amazing what you're doing and then we can dive into some more granularity.

Philippa Webb: Great. So our idea was to try to break down the silos between technology, law, students, practitioners, academics, companies, public and private sector, because we can't realise these fundamental rights and we can't advance justice if we don't think outside of these boxes. So the Oxford Institute's mission is to harness the power of AI to advance justice.

As you noted lawyers and the justice system don't move as quickly as technology does. That gap is becoming even broader. But there's another kind of gap, which is what we call the justice gap. Because according to the World Justice Project, only 10% of people who need a lawyer globally can access one and obviously that's an even smaller percentage if you look at vulnerable communities, if you look at women, if you look at global south countries. And we thought, what if technology in particular AI could play a role in trying to close that gap? And what role could we play in bringing together the right people to think about that?

So we're very fortunate that our technology partner is the Microsoft AI for Good Lab, and together with them, we've been thinking about and building tools to close that justice gap. But in addition to that, we're also thinking about how do we learn about what is happening in the world? How do we bring visibility to AI in the courtroom, to cyber litigation, to digital evidence, to deep fakes and so on?

So these are all projects that we work on at the Oxford Institute. What brings it together is we're thinking about both how AI can be a tool to advance justice, but where AI may be causing some injustices. How can we find roots to accountability?.

Susannah de Jager: That's a lot.

Philippa Webb: Yes.

Susannah de Jager: And you've put it beautifully. There's those two sides. So there's the huge opportunity and how we can expedite things both in the global south where there's very little access, but even in established legal systems where the backlogs, as we know, are huge. And then there's the risk side of things.

Philippa Webb: Yes.

Susannah de Jager: Because I would like us to finish on a more positive note, we're going to start with some of the risks. And you spoke about your focus being between states, between their citizens. And I know a lot of the work you're doing is looking at sort of the risks at a governmental and a citizen level from attacks. And I'd love you just to talk a little bit about that work.

Philippa Webb: Yeah, so 130 states to date have suffered cyber attacks and that number includes repeat attacks. We've seen it as a strategy that Russia is using against Ukraine in their conflict with cyber attacks on critical infrastructure, but we've also seen it in the UK in attacks on some government agencies and also on the private sector. There is no country that is immune from cyber attacks and associated disruption to critical infrastructure. The challenge now is that AI is supercharging the ability of groups to conduct these attacks. It can be relatively simple through AI tools to actually conduct these attacks and cover up your tracks.

So we've got law enforcement trying to keep pace with these developments. Whereas, AI is making it even more possible for very small groups, low resource groups to cause disruption. So until now, there's been a lot of focus on the political response to this. So naming and shaming, maybe sanctions, responsive attacks where that's possible and we really wanted to look at legal roots to accountability because we have a very well-developed international legal system. We have well-developed rules of international law that most states have accepted as binding that are relevant to cyber attacks and in particular to the protection of citizens affected by attacks on critical infrastructure.

So we published in October last year a policy brief on legal accountability for malicious cyber operations. We've been convening workshops and groups that cross both the private sector and government to discuss how we can implement some of the recommendations in practise. Part of that has been through a collaboration with the law school on the Oxford process on protections in cyberspace but also with tech companies that are very interested in how they can have other options for responding to such attacks.

I think the statistic is that Microsoft blocks 300 million attacks a day on its users. This is a common problem, a common issue where we're looking to expand the toolkit of responses.

Susannah de Jager: By its nature, a lot of this is cross-border. You were talking there about working within the UK. Is there a sense in the work you're doing that people realise there needs to be the kind of additional layer of protection or, it's obviously really difficult to work in concert, but...

Philippa Webb: Yes, so the UN is concerned with this and they're looking at mechanisms for accountability. So we're in conversation with the key players at the UN on this and we will be feeding into the solutions that hopefully they will be developing through that forum. But we're also, looking at a variable geometry. We're looking at different places where we can have impact through our research and recommendations. So we've been speaking to Europol, for example. We're speaking to different networks of law enforcement. We're obviously speaking with the private sector.

So you're absolutely right. It is cross-border and there's no point in doing something that's just within one jurisdiction. This is one area where I think we can actually have some really productive conversations because every state is facing this threat and every state wants to have more options for responding to it.

Susannah de Jager: You mentioned earlier the work that you're doing looking at deepfakes. I'd love to understand because this is such a kind of fast evolving space and, people are aware of it, but I'd love to hear it from a technical perspective what you guys are doing.

Philippa Webb: So our first policy brief was on legal roots to accountability for malicious cyber operations. In this October, we're going to be launching two new policy briefs that kind of mirror images of each other. So one is looking at the massive amounts of digital evidence that we're now seeing in prosecutions and other proceedings for atrocity crimes and the fact that they're not perhaps being as used by prosecutors, defence, and judges as you might expect. There's a sort of another gap between the collection of, this digital evidence from people's iPhones, from dash cams, from CCTV, from satellites, that massive amount, and then the tiny amount that we're seeing actually referred to in a meaningful way in eventual judgments.

So we're doing a study. We call this the sort of the positive approach of how can we use digital evidence, more effectively to in accountability for atrocities.

Susannah de Jager: And just to dig into that because it seems obvious, but I just want to ensure, is that because people are just too worried that it's not legitimate footage? Or is it because there's too much volume? Is it because they don't have a framework? Is it all three?

Philippa Webb: It's all three. So from the very beginning, you have problems of even storing this amount of information. So there's technical issues. Then managing it, which includes chain of custody, authentication. Is it admissible? But then there's just the more fundamental problem of trusting this information once it's gone through that process and judges feeling comfortable moving from the more traditional mode of having a witness on the stand who is examined and cross-examined to actually being able to look at maybe an AI generated recreation of an attack on a civilian object that's a combination of footage, satellite imagery, maybe time lapse as well, and being able to interpret that and feel comfortable with it in their eventual decision on the underlying violation.

So we're actually looking at the whole ecosystem and where there are gaps, that can be filled through protocols, through training, even just recommending better systems for the management of these huge amounts of data. Because it is heartbreaking in its own way for survivors of an atrocity, not only to have suffered the atrocity, whether it's their loss of property or loss of their family or injuries to themselves, that they have managed to capture that in some way on their phones or through photos and hoping that's going to be looked at and taken seriously in a accountability process. And at the moment, there's this huge chasm between that expectation and what is happening in practise.

Susannah de Jager: Gosh. Yeah, as a layperson, I would assume that if I filmed something and used it, that would be admissible and taken seriously. I think it's really important and do you think that the education piece, you touched upon it quite lightly there, but one hears the analogy in other domains where surgeons perhaps have done something the same way for a very long time and there's a minority who will adopt a new process that means they have to relearn it, even if the technology is fantastic. And I suppose that, that carries through here for judges perhaps?

Philippa Webb: Absolutely, and it's not even judges. Certainly in surveys have said that they're interested in training on this, but the same surveys have shown that they have great concerns about relying on evidence and this links to the other policy brief that may turn out to be a deep fake. So they're erring as the judicial mindset tends to be on the side of caution.

But I would say it's even a matter of training at the level of investigators and prosecutors and defence counsel who all need to be very comfortable with this type of evidence. Not just its authentication and admissibility and its assessment of its probative value. But also the more technical side of managing it, deploying it in an understandable way, linking it to the specific burden of proof that they may need to discharge. There's no global body, and there's very few national bodies, that provide that kind of guidance to law enforcement and prosecution and defence.

Susannah de Jager: Fascinating, and it's interesting because historically, eyewitness accounts are the top of the tree and yet they've been proven to be wildly out, often.

Philippa Webb: Exactly.

Susannah de Jager: And so it's interesting that something we're completely comfortable with and we all accept is is actually very error prone.

Philippa Webb: It is and that's where the great promise of AI may be for digital evidence because it's not just that you have the recording which is presumably real time and with the metadata and can be authenticated in this way. But that you can combine and really bring to life what happened in a certain location or to a certain community or person.

So it's the potential not to just hear on witness's testimony, even if you have other witnesses from the same time, but have these different types of evidence from satellite all the way down to the iPhone on the ground that can really bring to the court's attention what happened and who was responsible. And we're finding that potential, which very much already exists is not being realised in these very important proceedings.

Susannah de Jager: And you spoke about the other side.

Philippa Webb: So the other side is how easy it is to create a deep fake. And we just taught actually an optional course here at the Blavatnik School called the Tech and Justice Lab where we asked a number of the commercial chat bots to create certain deep fakes. At first they would resist saying, "I'm not allowed to do that or I can only do that if it's an authenticated image." But within two prompts, we were able to jailbreak it, and create deep fakes. Now those deep fakes might have a watermark in the corner that could be removed through a screenshot and cropping. This is not rocket science.

So it was just making real for our students and also for ourselves how easy it is to create these deep fakes. We've been convening and interviewing experts in the tech companies, but also in government and in law enforcement who are all dealing with floods of deep fakes in their work. I have to say we're launching our recommendations in October, but I have to say it's a very difficult subject because even Google, for instance, has come up with some really interesting ways of identifying deep fakes. Like they have this tool called Synth ID, which will be essentially an invisible marker of authentication in the image or the the document or the recording. But it's only going to work at the moment on Google devices.

So we're finding a very imperfect coverage in terms of response to this challenge of deep fakes. And we're also tracking through our AI Justice Atlas the judicial responses to deep fakes in courtrooms around the world and we're finding examples of audio recordings being used in custody disputes where you hear one of the parents saying horrible things about the child being deployed, to remove their custody in those proceedings. Judges have been just asking one witness if that sounded like the person and in using that as a method of authentication. Which clearly is deeply insufficient. In that particular case, it did turn out to be a deep fake audio recording.

Susannah de Jager: And that's pretty terrifying for you and I. Here we are putting our visual mark out there in the world.

Philippa Webb: Yes, exactly and it doesn't take much material to create a very good deep fake recording or video. So I'm sure you and I are already have sufficient material out there, which is a terrifying thought and that's why we're very motivated to work on modes of how we can authenticate and regulate this kind of material.

So we're having fascinating conversations. I have to say we're in the thick of it now. I'm sure we'll have a lot to say in October when we launch our report. But every meeting we have on this topic just confirms for us that there's a gap again in recommended responses to this. So we hope that we can contribute to closing that.

Susannah de Jager: You touched lightly upon the justice Atlas. I'd love you just to elaborate a little bit on what that tool is for anyone listening because it's a cool idea.

Philippa Webb: Our AI Justice Atlas, which can be accessed for free from our website, so techandjustice.org and then go to the AI Justice Atlas, is a globe tracking the use of AI and the regulation of AI in courtrooms around the world. We've got 15 jurisdictions that are covered and including the UK where you can click on the location and see are police force using facial recognition? Are they using predictive policing? Are we in the world of Minority Report already? How are our judges using AI? And what kind of rules do we have regarding it?

But what's even more fascinating for me is looking at jurisdictions like Brazil and Argentina and Columbia where judges are both of a different profile because in a civil law system, they are coming at a younger age into the judiciary. So they're judges in their 30s and 40s. They're teaching themselves coding and they're building their own agents. So you have a really vibrant culture of AI, for better or worse, when you think about the right to a fair trial, in these jurisdictions that's interesting to watch. China has said that it will have AI embedded in every courtroom by 2030 and it's already used very extensively, especially for first instance in small claims in China.

So we've got those 15 jurisdictions and we're adding 40 more by October and we're about to have a twin to our Justice Atlas, which will be the Cyber Litigation Atlas going back to the challenges of malicious cyber operations. That will track different categories of cyber attacks and uncover and reveal what responses have been of local law enforcement, but also joint operations, which we're finding states and companies are using more and more. So through both of these Atlases, we're hoping to bring visibility to a fast moving area.

Susannah de Jager: Fascinating. Absolutely fascinating. Philippa, I know that you were, touched upon it earlier, but just give us a little more detail on what you're doing to support judges because I know there's a tool being built out there as well.

Philippa Webb: Yes. So in line with our commitment to looking throughout the justice ecosystem, so not just creating a bottleneck further along, we're also looking at how we can help judges do their jobs more effectively. You look like at a jurisdiction like India that has under its current backlog of cases, 300 years of judicial time to clear that backlog.

You look at jurisdictions like Brazil, which is very dynamic in terms of AI, but still requires judges to decide seven cases a day just to stay on top of their backlog. And you realise that we need to assist judges in making their work more efficient so that they can dispense justice to more people, but also in a responsible way, because without the right to a fair trial, the bedrock of any safe and functioning society then we're lost.

So everything we do is very much grounded in how we can advance the right to a fair trial in the courtroom. So we've started off by building a prototype fair trial advisor, which is trained on the book that Amal and I wrote, The Right to a Fair Trial. Because frankly, judges don't have time to leaf through a thousand pages to find the thing they're looking for and it actually works very effectively as a chatbot. But in order to make it even more effective, our next step on that tool is to find a jurisdiction where we can integrate the local case law on fair trial and therefore make it a tool that's going to be readily usable for judges. They won't just have access to the international standards, but also to their national precedence in order to have instant advice in the courtroom that they can choose to deploy as they like.

Susannah de Jager: It's amazing. You've already said this, so I'm repeating the point. But it's mind-boggling when you think actually in this domain, how much is still being left to people's memory and their training and, as excellent as I'm sure so many judges are, these tools now exist and yet they're not being deployed. So I'm thrilled that you're working on these issues.

This Oxford Plus miniseries is brought to you by Equinox, Equitable Innovation Oxford. Equinox is a major regional partnership established by the University of Oxford to drive inclusive innovation and sustainable economic growth. Bringing together universities, local authorities, government, industry, investors, and community organisations, Equinox provides a unified voice for Oxfordshire, championing innovation that delivers economic impact alongside social value.

More than 80 organisations have signed the Equinox charter, committing to atract investment, remove barriers to growth, and ensure the benefits of innovation are shared more widely. By aligning Oxfordshire's world-leading research, enterprise, and talent, Equinox aims to unlock nationally significant growth, reinforce the county's position as a global innovation hub, and ensure that success creates skilled jobs, attracts investment, and delivers lasting prosperity across the region.

Drawing you again on the impacts on, as you said, on mental health, I'd love to understand a little bit more how we are deviating without perhaps enough focus from what's optimal for our children at the moment. There are these gaps from what we know and the impacts it has, what we have that's a sort of working live experiment, and what perhaps, those that study it such as yourself know are bad for long-term outcomes. I'd love to understand a little bit how you see the experiment we're running through those lenses.

So moving on a little bit from that justice focused area to some of your more grassroots work, I'd love to speak about some of the work that you're doing, particularly in Malawi, looking at giving women, and you touched upon it earlier, groups that traditionally would not have access to legal advice and I think the number that I saw this morning really shocked me that in the total of Malawi, there are 776 lawyers serving a population of 22 million. I'd love to hear what you're doing on the ground there to try and facilitate access to legal recourse for victims of domestic and sexual abuse.

Philippa Webb: Yes. So as you point out, Malawi is a legal desert. There's so few lawyers for the population and I have to say of those 776 lawyers, very few of them work pro bono on human rights issues. So we have chosen Malawi to be our kind of pilot jurisdiction for our tools designed to increase access to justice. Our partner in the Oxford Institute, our implementing partner is the Clooney Foundation for Justice and they've worked for 10 years in Malawi. And they have an amazing local partner, the Women Lawyers Association. A group of fabulous women lawyers working pro bono on all kinds of legal issues, but in particular, child marriage and gender-based violence.

Another reason we chose Malawi as a testing ground is because actually its laws on paper are perfect. It has prohibited child marriage. It has prohibited sexual violence against women. And yet 37% of girls are married before the age of 18. So we have, again, that gap between what is said on paper and what happens in practise. So we have been working in very close collaboration, we call it co-creation, with the Women Lawyers Association of Malawi. But also with their network of women and girls, first responders who are teachers, nurses, shelter operators, to really try to understand what kind of technological intervention would help expand access to justice.

We're actually in the middle of a sort of pilot testing at the moment with 500 people in Malawi on a legal tool, it operates on WhatsApp, it also operates on a 2G and low connectivity environments and we're also going to design it to operate offline, that provides legal information to women and girls about their rights regarding child marriage gender-based violence and economic abuse.

At the moment, we're building this with Microsoft's AI for Good Lab in English because the main language spoken in Malawi is Chichewa, and Chichewa does not exist in Google Translate. It does not exist in the UNESCO Rare Languages Project. It doesn't exist on DeepL. So in parallel with building this tool at First in English, we are also with the support of Microsoft building a Chichewa LLM that will be able to later make this tool accessible in the language that most people speak.

As we go along, we're learning so much from our local partners on the ground Because of literacy rates, it makes sense for the tool also to be voice to text, not just text-based. Also, geolocation is a feature that they said would be really useful as well. So this tool we're hoping is going to expand the reach of amazing organisations like the Women Lawyers Association to hopefully every woman and girl in Malawi to be able to understand their rights and connect with a lawyer who will fight for their rights if needed.

Susannah de Jager: And so you're looking after sort of both sides of it there. I also saw that there was a tool being built out that would then help the lawyers themselves expedite the processing, which again, we all know this, right? This is the hugest leap forward we've had in terms of time. That side of the implementation is also really important. I

Philippa Webb: It's really important because there's no point in expanding access to justice only for the bottleneck to move further along the chain. So we look at the whole process and we've been speaking to lawyers in Malawi and saying, " how can we help you do your job faster and more efficiently while maintaining its quality?"

And they've said essentially a form filling tool, which AI is very good at, and we're building this tool that will reduce the time for filling out a protective order form. Which is like a apprended violence order, restraining order, and the most effective tool they tell us for helping women and girls in, situations of sexual violence. We're reducing the time it takes to create that form for filing with the court from weeks down to a few hours.

We're very excited about this tool because this is one tool that we really see the potential for scalability. And our colleagues in Malawi are already asking us, could we build this tool for divorce applications, for custody claims for even inheritance disputes? So I think there's great potential for this to be helping lawyers and their clients to access justice more effectively.

Susannah de Jager: And I think that you've identified there a bottleneck and this is in Malawi where there's obviously a real issue, but there's bottlenecks everywhere. So we're recording this just weeks after the awful case in France of Lyhanna being murdered when there was an outstanding accusation, I think, against him and it transpires that there are 70,000 such outstanding accusations in France that relate to children that have been attacked in some way. And it's pretty horrifying to read things like that and realise that even in a country such as France, that there is a tail that long that isn't being expedited.

Philippa Webb: That's right and all the concerns and I think that cause some hesitancy in jurisdictions about incorporating AI such as hallucinations, are the biggest ones, that they're going to make up things that the footnotes will be incorrect and so on. There's technical solutions to that. When you work on a closed data set, as we do on our tools, hallucinations are minimised to a trivial extent. And then you always have the human review at the end as well to just fully account for that.

So I think once again, we see this drag in perception between " Oh, AI is going to make things up. We shouldn't be incorporating it. Its place isn't in the justice system" to the actual reality of where privacy concerns can be addressed, accuracy concerns can be addressed, also concerns about the sort of cost of AI tools as well rapidly being addressed through what they call small and light models now that use much less data. There needs to be the closing of that perception gap that could make a real difference, as you say, even in very developed judicial systems.

Susannah de Jager: And to that point around bounded models, I've also had the pleasure of interviewing Audrey Tang for this miniseries who, is the Ex-digital Minister for Taiwan and she was saying, similarly, that you can really overcome so many of these things and get them to protect the values, protect the system, the legal structure that you want them to. So I came away from that conversation and also from this already very hopeful about how we can control a lot of these perceived risks and then use it to strengthen and benefit both our justice and democratic systems.

You also using some of these tools to help journalists on the ground who have perhaps been arrested and are being accused, and again, often in a conflict zone or in a situation where their access. I'd love to just hear a little bit about that too.

Philippa Webb: So I think it's never been in modern times a more dangerous time to be a journalist. All the statistics from the committee to protect journalists or reporters without borders are showing that journalists are being targeted, including in democracies, more than ever just for doing their job.

We have a longstanding relationship, with the committee to protect journalists and when we co-created with them and asked where's your bottleneck? What would make a transformative diference to your work in protecting journalists? They said the first 48 hours during which a journalist is charged or arrested or detained is crucial. That's when they need legal representation. But not just any legal representation. They need someone who's ideally going to act pro bono or for affordable rates, especially in that critical early period who is familiar with the jurisdiction, who is admitted to the jurisdiction, and has a commitment to freedom of speech and journalistic rights.

We have been working with them to build a tool that will actually, in contrast to our Malawi tool, be global in reach. And as part of that we're developing guides to journalist rights, jurisdiction by jurisdiction, and we have the partnership with the Thomson Reuters Foundation for that as well. Which is incredible because they already have so much of the underlying material. We're also building out a network of lawyers who are willing to be on this tool as part of that network to be suggested to journalists at risk.

We're in the middle of co-creation on this tool now. And it's fascinating the differences between co-creating for a malawian woman or girl and for a journalist in the field. It has to be very tailored, but in different ways, and we're very excited about the potential of this tool, especially in these times where we see democracy and free speech under threat.

Susannah de Jager: Yeah. Really agree. I think it's really interesting that earlier in this conversation you said, "We didn't think it was another book."

Philippa Webb: Yeah.

Susannah de Jager: And I'm hearing you talk about all these kind of really action-based changes, that feels so important and tangible. What I think surprised me when I was looking into your work is that there isn't more being done already.

Do you think it's just that it's developing so fast that nobody's yet filled the space?

Philippa Webb: We've been surprised by that as well and we've been meeting with the top tech companies in this space with the top representatives of the judiciary in countries around the world and we've been saying, " Who's working on this? Who's thinking about how we use AI to advance justice?" And there are examples especially of very local startups that are working on digitization, digital case management, translation technologies. But it was really surprising that there wasn't this global approach to AI and justice. So that's why we're so busy in this space.

But we're very fortunate. What makes it possible is that we have this pro bono support from in - kind technical assistance of Microsoft and other tech partners, Thomson Reuters as well, HURIDOCS. We have a set of global law firms who are also providing us with pro bono assistance on all of the really interesting legal issues in this space, including do you have legal privilege if you get advice from a chat bot? Would they in a courtroom accept judges using a tool that's provided by a commercial company? Or do they want to have called judicial sovereignty over homegrown technology that's used in their domestic system?

So we don't feel so lonely, I suppose in that we've built up this partnership network. We've also, partnered with UNESCO on training for judges on the responsible use of AI through a massive online open course. So we're gathering a movement, I would say, because we've got a lot to catch up on. It seems like AI was first deployed and continues to be deployed to great effect in science and medicine. You've already got your huge data sets ready to analyse. We're seeing it in finance for the obvious profit imperative there. We're even seeing it in entertainment, again, profit is driving that and just the fascination of popular culture. So we need to bring up the profile of justice as well. Justice is also a public good that we need to protect and advance.

Susannah de Jager: And so it sounds like you've got amazing supporters, but if people are listening to this, which we hope they will, what do you need more of?

Philippa Webb: That's a wonderful question to ask. So we are always looking for local partners because, in the justice sector, whether they're legal aid or whether they're in tech injustice or even in in academia, because we're starting off with our, women and children's tool and our lawyers tool in Malawi. But we would love to spread this to other jurisdictions. But we can only do that, because we're a small research institute, we can only do that with amazing local partners on the ground. So we're always looking for people who are really familiar with the jurisdiction who perhaps have connections with the judiciary.

We're also looking for, in another way, technologists who may have built maybe an amazing tool in the healthcare sector that's a chat bot that has many similar features, privacy, natural language processing, geolocation, and we would love to learn from that for our tools in the justice sector.

Susannah de Jager: I'm making a list of people to introduce you to.

Philippa Webb: Oh, excellent. Another thing on a really basic level is that a lot of what we do is only possible if the laws are already digitised and what we're finding is that even in the UK, not all of our laws and cases are digitised. In Malawi, only 30% of their laws are digitised. So it's very difficult to train a tool that's going to be comprehensive and effective if you don't have that underlying material. So there's just also the very essential foundational work of digitising these laws.

Susannah de Jager: Yeah. Fascinating. Looking forward, where do you hope the institute and indeed some of these issues will be in sort of two, three years?

Philippa Webb: So we hope that the Institute will become the gold standard for AI and justice work. That people will turn to us for best practises in this area and that we will be able to share, practises not just in Malawi but we hope by then that we would have scaled to maybe 20 more countries with our tools.

It's interesting. I don't think we're going to expand our toolkit in a way. I think we will be focusing on the rights of women and girls, on the rights of journalists, on how to help judges. But what is going to change is tailoring those tools to each jurisdiction. And then through our Cyber Litigation Atlas and our AI Justice Atlas developing policies and recommendations to assist governments around the world in dealing with this fast moving and potentially very impactful technology and later this year, we will be launching a programme for students to undertake paid internships with the Institute. So we would love to hear from law students, from philosophy students, from computer science students the interdisciplinarity of Oxford is a real gift in this area and to learn from them as well as share our work with them.

Susannah de Jager: Amazing. Philippa, thank you so much. I've really enjoyed this conversation.

Philippa Webb: It's a pleasure.

Susannah de Jager: Thank you for listening to this episode of Oxford Plus, hosted by me, Susannah de Jager If you wanna keep up with all things Oxford Plus, visit our website, oxfordplus.co.uk or sign up for our newsletter on Substack.

Oxford Plus is a podcast produced by Story 94.