Speaker 1: When I was 10 years old, I visited a prison in India. It was dark, smelly, and really scary for a kid. But what I remember from that visit is meeting a 10-year-old guy, a kid who should have been stuck in jail for two years, who had instead been there for five. Again, he was 10. There was no reason for him to be there. But the legal proceedings around his case made it such that he was stuck in a system where he had no empowerment. He, despite what the law said, had been excluded from the justice system, like many others, and therefore had been taken advantage of. Throughout the years, I visited more and more countries, from India to Cambodia to Myanmar and Tibet, each time meeting with government officials and local lawyers to ask them why I kept hearing stories like this and why the state of access to justice isn't where it should be. Very quickly, I came to learn that the problem of access justice is not due to a lack of empathy, but due to a lack of resources. Can't afford a lawyer? Well, the government would love to give you one, but that costs money, which often isn't there. Don't understand your case information? Well, getting you a legal expert that could go in and explain it to you has billable hours involved, billable hours which we can't necessarily pay. Want to make a phone call to your family and tell them that you're not coming home tonight because you got locked up in jail and you're not going to be back for a long, long time? Sure, go wait in line, but the line for the phone is 500 people long, and you have not earned the right to skip it. These issues sound like things that you'd only experience in emerging economies, but even in countries like the United States, much better funded than the other nations I listed, 86% of civil issues faced by low-income Americans are met with either no legal aid or legal aid that is insufficient. Clearly, the justice system is broken at all levels, but it was always bothered me the most is the lowest level, those who are taken directly from their homes and put into jail cells, skipping the courtroom and the chance to defend themselves in between. The dream of an inclusive justice system that leaves nobody behind is as old as justice itself. And in my limited years working on the problem and the much, much, much research that exists, it's clear that access to justice is about access to counsel. And to provide that, there are really two very simple avenues. The first is to have an army of lawyers much bigger than what we currently have ready to go down and volunteer their time, to work not in the private sector but the public sector, make much less money but help those in need. The second is to enhance the ability of the existing lawyers to create infrastructure that can automate their work and elevate the things that they can do to help more people. The first we're nowhere near achieving on a global scale, but the second is where AI comes in. The idea of AI taking over the legal profession is a long and storied one that has been very well documented but even better funded, and for good reason. On paper, it sounds like a perfect marriage, an industry rooted in analyzing vast amounts of data to make complex decisions with intricate rules, exactly the kind of thing that artificial intelligence seemingly excels at. Today, we're closer to this ideal than ever. Companies like Do Not Pay are deploying legal bots that are able to write up entire police reports or lawsuits or fight your parking tickets with the click of a button for pennies on the typical human lawyer's dollar. And with the widespread accessibility and integration of models like GPT-4, which, as you may have heard, score in the 90th percentile of all bar examinees, people are now asking themselves, once again, are lawyers going to be replaced by AI? And again, the answer is no. At least, not yet. Here are three big reasons why. Number one, and probably the biggest issue, is that AI simply isn't good enough yet. While AI may be able to pore over troves of legal documents and make complex conclusions, citing laws as its reasoning, the reality is that it has no idea what the laws it's citing actually means. This lack of expressive power severely limits its ability to make decisions under uncertainty, but more than that, is deeply concerning when trying to uphold certain standards. We can tell an AI, for instance, not to be biased, but as many of you, I'm sure, have seen, this doesn't work very well in many cases, from sexist hiring bots to rehabilitation bots, which predict things like black people being five times more likely to commit a crime, stats that don't track in real life. So, how can we get to a place where AI is able to help lawyers in what they do? Well, in addition to not being good enough, we also have the culture issues, right? We go into a court and we swear an oath of truthfulness, and we're not good enough to accept that for computers. And more than that, we have issues of liability and trust. If an AI gets something wrong, if an AI gets something wrong, then who is to blame? The second set of issues, the cultural ones, are things we can't ignore. And the third set of barriers keeping AI from taking over the legal profession are barriers around the human factor. Of course, being able to make legal conclusions and read documents is important, but any lawyer knows that that's only half the job. The other half, from courting clients, maintaining relationships, and going in and convincing a jury of human people of the emotional rectitude of your case is important. And until artificial intelligence can get to a level of emotional acumen, physical presence, and ability to instill trust in the people around it, then human lawyers can hold safe to their jobs. So, we see that human lawyers are not going to be replaced by AI. Rather, AI is a powerful tool to assist them. Do Not Pay may be making these systems which, with a team of 16 people, is able to assist over 200,000 active clients, that's over 12,500 clients a person, but they're still under fire for issues of accuracy, issues of bias, and issues of certification with their bots. So, for now, AI is out-competed by their human counterparts. But what about people who don't have access to the human counterparts? People like the man I met in India who are currently left behind and have no resources at their disposition. This is where things get fascinating, and where we see the asymmetry on the effect of AI on the justice system. While those who already have access to competent legal counsel will experience mainly boost in efficiency, those who have nothing will be raised to a certain baseline that they have never before seen. For them, it can act like a lifeline. Earlier, I said we needed one of two things to make access to justice a reality for all, that being more lawyers or proper infrastructure. AI is special in the fact that it can provide this infrastructure at unprecedentedly low costs and incredibly high scale. We can employ many experts to automate lots of the work that lawyers do for people that wouldn't have a lawyer in the first place. And that's where we see huge return on investment. Governments around the world are already racing to try and integrate this in their countries. In LA, for instance, the state government is trying to help integrate a Q&A bot, which helps people with their legal filings. Since introducing this, the success rate of these filings has gone up exponentially, showing how providing expertise through automated means can help truly increase access to justice. Additionally, these systems are, again, so low cost and easy to produce. And I am the proof of it. This summer, exactly a decade after I first went to India, I went back with another student and worked on a fellowship to solve our own piece of the injustice puzzle, access to information. 70% of all prisoners in India are illiterate, meaning that although they are allowed to read their information and are given it on pieces of paper or on screens, they have no idea what's going on with their case. They, again, cannot read. In addition to causing immense anguish, it causes a huge amount of workload for lawyers, paralegals, and experts, which often don't exist. Our solution to this is deceptively simple. A voice bot that talks to them, thereby bypassing the illiteracy issue, which is able to read their information and then answer questions dynamically in a conversational way. What's the status of my case? What's my charge? What's going on with my legal proceedings? It can go through their document and read it out to them. What's important to note here is that we're not automating an entire lawyer's work, just one piece of it, but it's a very time-consuming piece and one that we can deploy at huge scale for people at very low costs. It sounds crazy, but a system created by just two college students is now geared to help hundreds of thousands of illiterate prisoners. Sure, we had support of a fellowship and a nonprofit on the ground, International Bridges to Justice, but the resources invested were so minimal that it would probably blow your mind to hear the numbers. And meeting with government officials, from members of the Supreme Court to members of Parliament to directors of jails, we felt a huge enthusiasm within them to implement these solutions and truly solve these gaps in the justice system. Currently, AI is making a legal landscape where the lack of resources is no longer a proper excuse. Since AI can so vividly help and plug these problems with limited funding, then it's now an issue of a deficiency in innovation. This paradigm shift is huge and means that in the right hands, we could truly help plug and bring a legal baseline that has never before been seen. Just as a hammer cannot build a house, AI may not be able to replace all lawyers, nor may it be able to completely fix the justice system or provide justice to all. It can, however, in the hands of the right people with the right caution and the right time, help bring a baseline that many people do not face and truly create a more equitable and just justice system all around the world. Thank you.