Earlier this month, the Legal Services Corporation (LSC) kicked off 2019 with the Innovations in Technology Conference, an event that combined some of the brightest minds in legal tech to devise solutions aimed at tackling the deficit of civil legal services for low-income Americans.
Founded in 1974, LSC is a pioneer in the access to justice movement and the nation’s largest funder of civil legal aid services for low-income individuals, providing grants to legal service providers across the country. LSC helps individuals whose household income is at or below 125% of the federal poverty guidelines with a focus on expanding the availability of civil legal services from family law, to housing issues, consumer issues, employment issues, and disaster response for low-income individuals. While the LSC largely provides assistance through clinics and self-help materials, it also focuses on connecting available lawyers with needy individuals: a mission that was the centerpiece of the 2019 conference.
This year’s LSC Conference featured a broad range of topics from access to justice for senior citizens to building AI infrastructures, but in this article we will focus on one particular topic: A panel discussion entitled, Building the Legal Help AI Infrastructure: Taxonomies, Datasets, and Classifiers. Led by David Colarusso (Legal Innovation & Technology Lab), Erika Rickard (The Pew Charitable Trusts), Margaret Hagan, (Legal Design Lab), and Sarterus Rowe (Legal Service National Technology Assistance Project), the panel’s discussion centered on how to create an AI infrastructure to connect lawyers to clients in need of civil legal services.
Developing an AI Infrastructure: Perspectives from Erika Rickard and Margaret Hagan
Throughout their presentations, the panel’s focus centered on the classification of legal terms to, as Erika Rickard said, “chip away at the access to justice crisis in our country” and to build a legal assistance AI infrastructure. Rickard started the panel discussion with an introduction about why the missions of groups like LSC matter. “There is an access to justice crisis in this country,” noted Rickard. “The word crisis implies something new and unpredicted, and unpredictable, and that we are urgently solving, and unfortunately, that’s not the case. It has become the norm.”
Rickard went on to describe that for decades, millions of Americans have attempted to navigate the legal system on their own. What groups like The Pew Charitable Trusts are doing in conjunction with LSC is finding new tools to address these issues, such as using machine learning to more accurately and quickly connect low-income clients to legal information and services. As Rickard explained, the access to justice gap isn’t just about more people needing help and a shortage of lawyers – it’s about lawyers who are able and willing to help, but who don’t know how to connect with the needy client population. What’s clearly necessary is an intersection between attorneys who want to help, clients who need help, and legaltech tools to help bridge that gap.
“It’s a big challenge,” says Rickard, but groups like Pew and LSC are up to the task. Rickard’s introduction dovetailed into a compelling presentation by Stanford Law School’s Margaret Hagan, a trailblazer in legal design and innovation. Hagan began her presentation with the question of how, in the “nascent moment” of new developments in artificial intelligence, can we ensure that the systems we use to connect lawyers and clients are “interoperable, standardized, and consistent.”
In Hagan’s presentation, she discussed the basics of the National Subject Matter Index (NSMI), which was initially created as a laundry list of specific civil legal aid issues. Written by lawyers, for lawyers, the NSMI initially had no governance structure or interactive, user-friendly interface. It was essentially a static list, and as Hagan explained, it was not a good fit for integrating with AI technology, due to a lack of consistency or normalization across data sets. What Hagan and other legal tech innovators have accomplished is to fine tune a machine learning model to take in text and pull out the most salient, searched legal issues.
To illustrate this, Hagan pointed to a standard landlord-tenant scenario:
TENANT: “I have been living in my apartment for nearly 2 years now. Shortly after I moved in, there was a leak in my ceiling that damaged my mattress beyond repair. My landlord grudgingly paid for me to replace it though angrily told me he would replace it for $50. He is an angry cheapskate but told me it was fixed and we had no issues. Fast forward to yesterday when the same pipe burst. Soaking my mattress monitors and keyboard. I am looking at $500-$600 to replace.”
Hagan then posed the question: Do you see a legal issue around eligibility in this post? The answer is, clearly, no. As Hagan detailed, this example demonstrates the need to devise a consistent infrastructure across a wide range of AI projects to better reflect not only how people talk about their legal issues, but also how lawyers categorize them. In other words, the NSMI needs to be rewritten to reflect a taxonomy that not only allows lawyers to categorize information, but that also helps the end users to understand it, while interacting effectively with various AI systems. However, the biggest challenge with the NSMI and “NSMI 2” is keeping the lists up to date, as Hagan noted that multiple updates throughout the year would be needed to keep the system as effective as possible.
UniCourt’s Role in Closing the Access to Justice Gap through Tech
UniCourt consistently works to improve search functions to normalize and create common case types, case statuses, entity names, and issues across jurisdictions at the federal and state level. This is no easy task, as most every court system speaks its own language, and not all court systems provide the same level of court data publicly. Streamlining searches by case type also has direct corollaries to the NSMI project. Much like the Legal Service National Technology Assistance Project and other groups, UniCourt is developing methods to make large data sets more effectively interact with machine learning technology to better connect consumers with the resources they need.
As Erika Rickard noted, the access to justice gap involves not a shortage of attorneys, but rather, a challenge in connecting skilled, service-minded attorneys to needy clients. There are numerous lawyers who are willing and able to do their part to improve access to justice, but their emotional and intellectual capital is wasted if we cannot adequately connect them to the growing client population in need of their help.
With UniCourt, lawyers looking for clients in need of free or affordable access to legal services can search for self-represented parties who are defendants in landlord-tenant cases, collections matters, family disputes, and more. Using UniCourt’s party type search filters, along with case type and jurisdiction filters, lawyers can quickly narrow from tens of thousands of cases in their state, to the most relevant cases at their local courthouse where self-represented litigants are wading through legal issues on their own.
Now imagine this a scale. Legal aid organizations can leverage UniCourt to automate the process of finding and assisting self-represented litigants, by scheduling automated searches on a daily, weekly, or monthly basis to find all pro se parties for specific case types the organizations have mandates to handle. Taking this one step further, if these legal aid organizations have a matter management system, they could also utilize UniCourt’s Legal Data APIs to directly input case information into their systems and receive notifications when new cases fit their criteria. Moreover, if these legal aid organizations have an active registry of pro bono attorneys broken down by practice area, they could then automatically assign new cases to attorneys with the requisite expertise.
At UniCourt, we are dedicated to making court records organized, accessible, and useful. We are committed to connecting attorneys, businesses, and consumers to the records they need and enable them to tap into the mountain of court data generated everyday. To that end, we welcome partnerships with nonprofits, state, local, and nationwide legal aid organizations, law school clinics, and others interested closing the access to justice gap.
Contact us if you are interested in learning more about UniCourt’s products and services, and if you are interested in accessing bulk court data through our Legal Data APIs.