Access to Justice and AI
This week on “Waking Up With AI,” Katherine and Anna discuss the potential for AI to increase access to justice, while flagging some risks and ethical concerns around the use of AI in law.
- Guests & Resources
- Transcript
Katherine Forrest: Hey, good morning everyone. I'm Katherine Forrest.
Anna Gressel: And I'm Anna Gressel. We're your hosts for “Waking Up With AI,” a Paul, Weiss podcast.
Katherine Forrest: And Anna, this morning I'm actually going to confess to you that I have a Diet Coke in my hand, which I'm holding up. Nobody else knows that we're actually on Zoom with each other, but we do. And I felt like I needed something sort of really sort of punchy this morning, different from coffee, because I just got back from a judicial conference in Chile. I speak at a, actually at a lot of their judicial conferences.
Anna Gressel: I know you do, and I mean, I know you used to be a federal judge, but does our audience know that?
Katherine Forrest: I’m not sure actually how much background we've given them, but what I do is I combine my years of judicial experience with my passion for AI. And it's really gratifying because I'm able to, I hope, bring a perspective to the AI issues that has a sort of a novel bent. And on that note, I wanted to talk today a little bit about something I get asked about quite frequently at judicial conferences: access to justice issues.
Anna Gressel: I think that's such a terrific topic, Katherine, and there's so many different ways that generative AI has the potential to impact access to justice, just like the legal industry. But with many things around AI, I think we should start off with a few preliminary points for our audience. The first is, what do we even really mean by access to justice? And second, how does generative AI possibly impact that?
Katherine Forrest: Anna, the phrase “access to justice” can mean really a lot of things, and people often don't define it. They just sort of use the term. But here, what I'm referring to is the ability to use generative AI tools to increase the ability of people and organizations to navigate the judicial system and literally obtain increased access to the decision-making and dispute resolution processes that the judicial system has to offer.
And when you're talking about GenAI or generative AI, let me give our audience a sense of what's at issue. Many of the GenAI tools that we're familiar with these days, these large language models like ChatGPT, Bing Chat, et cetera …
Anna Gressel: Yep.
Katherine Forrest: Those models allow a person...
Anna Gressel: Just like a regular person.
Katherine Forrest: Right, just a regular person, to download an app on his, her, or their phone and they can ask it questions, but they can do more than that. They can also draft documents, upload documents into the app and ask the app to summarize the documents for them or tell them what the documents mean in a language that they might find more accessible or just accessible ab initio.
Anna Gressel: Right, so just to throw in an example, you can put a document into ChatGPT or a different program and you can actually ask it to translate it into another language like Spanish, actually hundreds of languages these days.
Katherine Forrest: Right, but you can also have it summarize what the document is saying in non-legal jargon or language that's accessible to a person at any particular educational level from their own language to their own language. So for instance, it can be in the English language starting out in non-accessible legal jargon and be translated into the English language in non-legal jargon.
And that functionality—translating legal language—allows a person who's unfamiliar with certain words or even how to understand certain instructions about what to do next, to read or hear that document and make it all that more accessible.
Anna Gressel: I mean, Katherine, I'd love to hear your views on this. You're a former federal judge, and you had all kinds of folks in your courtrooms. It must be really empowering and important to help people understand what can be a really confusing and scary legal process. And I just want to go back to one other point that we can talk about, about how generative AI tools can enhance access to justice, and that's actually how people use tools.
Katherine Forrest: Right, like a ChatGPT.
Anna Gressel: Yeah, and they can actually use it on their phone to have legal questions answered. So for example, if X, Y or Z happen to me, do I have a claim?
Katherine Forrest: There are some risks.
Anna Gressel: Definitely. I mean, I think right now these tools still have quite an inaccuracy rate and that's one pretty significant risk that people, both lawyers and non-lawyers, should be aware of.
Katherine Forrest: Right, there are these things called hallucinations. Those are those inaccuracies that you've mentioned.
Anna Gressel: Right, and there are real differences between the tools as well. Some of these tools have monthly fees and they get access to more updated information. Some don't. And so you're not even really getting the most recent use of the tool, or the most recent news from the tool unless you're actually paying that fee.
There are also confidentiality issues. So some functionalities on apps you just download on your phone don't actually have the kind of confidentiality you might want from an AI tool. You have to be really careful. I mean, people in the public have to be really careful to not put in names or birth dates or addresses, stuff that could identify a real person.
Katherine Forrest: But even with these limitations, there are really some serious benefits. And as the tools get better and better, and people can actually access tools that are made specifically for the legal profession, that are fine-tuned for the legal profession, the benefits will become even greater.
Anna Gressel: So how do you think about that, Katherine, for not-for-profit organizations that are doing legal work?
Katherine Forrest: Well, they'll also be able to utilize, in fact, the more specialized legal tools to take on even more cases or matters or to give more advice more quickly because what these tools do is they speed along a series of tasks.
Anna Gressel: You know, Katherine, I've actually used some of these tools as part of a beta. Harvey's one tool that we use sometimes, and it can draft documents in seconds. So you can see how a lot of legal work would actually get done pretty quickly with AI assistants.
Katherine Forrest: Oh, for sure.
Anna Gressel: So let's talk a little bit about the ethical obligations that lawyers have if they decide to use AI tools in this way.
Katherine Forrest: Well, the ethics of using generative AI in connection with legal work is really worth an entire episode that we'll probably do at some point, Anna. But it's just to sort of hit a couple of highlights here — I'm just going to really bullet-point them off — which is that:
- You always have to read the cases that are cited by any generative AI tool to you, because there could be hallucinations, they could not be right.
- Sometimes there are courts that have certain requirements in terms of disclosure. You may have to tell the court if you're using GenAI. Some courts don't even allow the use of GenAI. So you've got to really understand that, along with
- The duty of confidentiality and making sure that you're not overstepping the confidentiality obligations you have.
Katherine Forrest: All right, so our time went really fast today. It's a short podcast, but we like to get a little information out there. And with that, I'm Katherine Forrest.
