Artificial intelligence poised to hinder, not help, access to justice – Reuters

April 25 (Reuters) - The advent of ChatGPT, the fastest-growing consumer application in history, has sparked enthusiasm and concern about the potential for artificial intelligence to transform the legal system.

From chatbots that conduct client intake, to tools that assist with legal research, document management, even writing legal briefs, AI has been touted for its potential to increase efficiency in the legal industry. It's also been recognized for its ability to help close the access-to-justice gap by making legal help and services more broadly accessible to marginalized groups.

Most low-income U.S. households deal with at least one civil legal problem a year, concerning matters like housing, healthcare, child custody and protection from abuse, according to the Legal Services Corp. They dont receive legal help for 92% of those problems.

Moreover, our poorly-funded public defense system for criminal matters has been a broken process for decades.

AI and similar technologies show promise in their ability to democratize legal services, including applications such as online dispute resolution and automated document preparation.

For example, A2J Author uses decision trees, a simplistic kind of AI, to build document preparation tools for complex filings in housing law, public benefits law and more. The non-profit JustFix provides online tools that help with a variety of landlord-tenant issues. And apps have been developed to help people with criminal expungement, to prepare for unemployment hearings, and even to get divorced.

Still, there's more reason to be wary rather than optimistic about AIs potential effects on access to justice.

Much of the existing technology and breakneck momentum in the industry is simply not geared toward the interests of underserved populations, according to several legal industry analysts and experts on the intersection of law and technology. Despite the technology's potential, some warned that the current trajectory actually runs the risk of exacerbating existing disparities.

Rashida Richardson, an assistant professor at Northeastern University School of Law, told me that AI has lots of potential, while stressing that there hasnt been enough public discussion of the many limitations of AI and of data itself." Richardson has served as technology adviser to the White House and Federal Trade Commission.

"Fundamentally, problems of access to justice are about deeper structural inequities, not access to technology," Richardson said.

It's critical to recognize that the development of AI technology is overwhelmingly unregulated and is driven by market forces, which categorically favor powerful, wealthy actors. After all, tech companies are not developing AI for free, and their interest is in creating a product attractive to those who can pay for it.

Your ability to enjoy the benefits of any new technology corresponds directly to your ability to access that technology, said Jordan Furlong, a legal industry analyst and consultant, noting that ChatGPT Plus costs $20-a-month, for example.

Generative AI has fueled a new tech gold rush in "big law" and other industries, and those projects can sometimes cost millions, Reuters reported on April 4.

Big law firms and legal service providers are integrating AI search tools into their workflows and some have partnered with tech companies to develop applications in-house.

Global law firm Allen & Overy announced in February that its lawyers are now using chatbot-based AI technology from a startup called Harvey to automate some legal document drafting and research, for example. Harvey received a $5 million investment last year in a funding round, Reuters reported in February. Last month, PricewaterhouseCoopers said 4,000 of its legal professionals will also begin using the generative AI tool.

Representatives of PricewaterhouseCoopers and Allen & Overy did not respond to requests for comment.

But legal aid organizations, public defenders and civil rights lawyers who serve minority and low-income groups simply dont have the funds to develop or co-develop AI technology nor to contract for AI applications at scale.

The resources problem is reflected in the contours of the legal market itself, which is essentially two distinct sectors: one that represents wealthy organizational clients, and another that works for consumers and individuals, said William Henderson, a professor at the Indiana University Maurer School of Law.

Americans spent about $84 billion on legal services in 2021, according to Henderson's research and U.S. Census Bureau data. By contrast, businesses spent $221 billion, generating nearly 70% of legal services industry revenue.

Those disparities seem to be reflected in the development of legal AI thus far.

A 2019 study of digital legal technologies in the U.S. by Rebecca Sandefur, a sociologist at Arizona State University, identified more than 320 digital technologies that assist non-lawyers with justice problems. But Sandefur's research also determined that the applications don't make a significant difference in terms of improving access to legal help for low-income and minority communities. Those groups were less likely to be able to use the tools due to fees charged, limited internet access, language or literacy barriers, and poor technology design.

Sandefur's report identified other hurdles to innovation, including the challenges of coordination among innumerable county, state and federal court systems, and "the legal professions robust monopoly on the provision of legal advice" -- referring to laws and rules restricting non-lawyer ownership of businesses that engage in the practice of law.

Drew Simshaw, a Gonzaga University School of Law professor, told me that many non-lawyers are "highly-motivated" to develop in this area but are concerned about crossing the line into unauthorized practice of law. And there isn't a uniform definition of what constitutes unauthorized practice across jurisdictions, Simshaw said.

On balance, it's clear that AI certainly has great potential to disrupt and improve access-to-justice. But it's much less clear that we have the infrastructure or political will to make that happen.

Our Standards: The Thomson Reuters Trust Principles.

Opinions expressed are those of the author. They do not reflect the views of Reuters News, which, under the Trust Principles, is committed to integrity, independence, and freedom from bias.

Thomson Reuters

Hassan Kanu writes about access to justice, race, and equality under law. Kanu, who was born in Sierra Leone and grew up in Silver Spring, Maryland, worked in public interest law after graduating from Duke University School of Law. After that, he spent five years reporting on mostly employment law. He lives in Washington, D.C. Reach Kanu at hassan.kanu@thomsonreuters.com

Original post:

Artificial intelligence poised to hinder, not help, access to justice - Reuters