Features

Technology as the future of access to justice

 

by Kimberly Paulson   |   Michigan Bar Journal

Accessibility to legal services takes many forms. Before the internet, the primary obstacles to accessing legal services were geography and the ability to travel to a legal aid office or courthouse. Websites and email changed that, and smartphones have further bridged that divide. Now, people can access legal information from the comfort of their homes. So, why are so many unserved?

The limited resources of legal aid providers is one obvious answer. If humans are required to serve clients’ legal needs, staff constraints will always be a broken spoke in the wheel of justice. Courts, legal aid providers, and pro bono attorneys are also evolving to try to meet indigent clients where they are, both figuratively and literally.

Many believe that the answer is not just more technology but using existing technology in a way that better meets the needs of both legal aid providers and underserved populations — innovations that not only reduce the level of human interaction required, but provide a more familiar, simple experience for those involved in the justice system. The COVID-19 pandemic brought some of these technologies to the forefront as courts and legal aid providers scrambled to better serve individuals from afar. As a result, we can picture a future where technology narrows the justice gap.

ZOOM ZOOM

Since the start of the pandemic, lawyers have gained a greater familiarity with online meeting applications such as Zoom and Microsoft Teams. Remote court hearings, client interviews, and staff meetings suddenly became easier and more efficient. These applications hold great potential for increasing access to justice. Low-income clients often miss attorney meetings and court hearings because they cannot leave work, can’t afford childcare, and/or have no money for bus fare or parking. If the client could simply download a free app and attend the scheduled event remotely, the number of missed court dates would undoubtedly drop. Data collected during the pandemic will shed more light on the accuracy of that statement, but the fact remains that the traditional model of in-person appearances and meetings is an inherent disadvantage to low-income clients. Online meetings are one key to greater participation in the justice system.

While online meetings have the potential to increase access for some, they can also create barriers to access for others, including people with disabilities.1 Not everyone has access to a reliable internet connection, a computer, or a smartphone with a strong enough signal and a sufficient data plan to participate in remote hearings. The digital divide is still very much a problem in Michigan;2 to fully realize the potential benefits of remote hearings, courts and others in the justice community must work to provide litigants with the proper tools to fully participate in remote hearings and meetings.

MAGIC FORM-ULA

Automated document assembly software has emerged as an important way to help unrepresented parties with the daunting task of selecting and properly completing court forms. The software, which may be included as part of a comprehensive law practice management system or a stand-alone tool, auto-populates customized templates of legal forms and documents.3 The data input into the templates may come from a database linked to the software or may be collected from the user, often through online interviews using branching logic, a concept discussed in more detail below. Using the information obtained during these question-and-answer sessions, the tools can determine whether the litigant is eligible for the relief they want, find the appropriate court form, and automatically fill out the form with the user’s information.4 Once the forms are completed, the user can then print them out and file them with the court. Tools like these are crucial to providing those who cannot afford attorneys real access to the courts.

Michigan leads the way in this area. The Michigan Legal Help (MLH) program operates a comprehensive website that includes 50 do-it-yourself (DIY) tools using this technology.5 MLH also supports self-help centers across the state in courthouses, libraries, and other locations, many of which are staffed with individuals who can help with the DIY tools. Legal aid providers and pro bono attorneys can use these tools to complete forms for their clients, making the process more efficient and giving pro bono attorneys the information and confidence needed to help people in areas outside of their expertise. Adoption of statewide court forms and document assembly solutions by all 50 states and U.S. territories — the ultimate goal — would open courthouse doors to countless unrepresented individuals and increase the efficiency of legal aid providers nationwide.

IT’S NOT JUST ROBOT BUTLERS

Artificial intelligence (AI) is increasingly becoming part of our everyday lives. Just ask Alexa. But despite the mental image of robot butlers, most AI used today is not flashy. It is embedded in computer programs and websites to make tasks faster and easier. AI is also being used in not-so-obvious ways to increase access to justice.

Natural language interpretation (NLI), also known as natural language understanding (NLU), is one function of AI that is particularly relevant to access to justice. Computers can only understand what they are programmed to understand. The varying cadence, vocabulary, and nuances of the natural way people speak and write have always posed challenges to programmers. Computers can typically recognize certain terms or phrases typed into a search engine (e.g., “eviction” or “expunge my felony”) through a matching algorithm, but problems arise when users do not use the anticipated keywords or spell them differently. Slang, misspelled words, and colloquialisms are often incompatible with conventional matching algorithms, resulting in the inability to access relevant information for the user.

That’s where NLI comes in. NLI technology can be trained to interpret the natural use of language by using familiar words and phrases to find answers to the user’s legal questions.6 Instead of the user slowly working through a logic tree and hopefully selecting the correct options along the way or typing in a query that produces no results, AI with NLI abilities should take the user directly to the resources they need, resulting in a more efficient search and more accurate results.

One example of NLI is an AI issue spotter aptly named Spot, created by Suffolk University’s Legal Innovation and Technology Lab (LITL) and provided to non-profit and government organizations at no cost. Spot is an application programming interface; a building block third parties can use in its applications or on its websites.7 Spot offers NLI capabilities that then become part of the final program or application. AI with NLI capabilities is ideal for incorporating into an online legal resource directory such as michiganlegalhelp.org.8 In fact, according to MLH Director Angela Tripp, the program is looking to incorporate Spot or a similar tool as part of its 2022 website redesign.

NLI is especially useful because it can learn to understand users’ intent.9 In creating Spot, LITL used publicly available historic questions posted on Reddit’s legal advice forum to train its AI to recognize how people use natural language to seek legal information.10 LITL continues to use crowdsourced data obtained from the Learned Hands online game11 and organizations using Spot to continuously improve the product. As explained by LITL Director David Colarusso, the more Spot is used, the better it becomes. It’s a collaborative effort in which current users help make Spot’s NLI better for future users.

The use of symbolic AI (SAI) is also important for access to justice because it uses logic and reasoning to reach a conclusion or accomplish a task, enabling machines to complete specific functions typically performed by humans.12 Think of SAI as a flowchart using “if x, then y” logic, a concept often referred to as branching logic.13 It allows computers to ask further questions and reach conclusions based on the user’s responses to previous questions. In the context of legal services, SAI can automate routine tasks or processes, making them more efficient and less time consuming. SAI is used in the document assembly solutions discussed above.

SAI can also be used for client intake because it understands which questions to ask at each step based on the information previously provided by the prospective client. It can arrive at conclusions after evaluating input information — including whether the prospective client qualifies for services and to which attorney or group that person should be referred — and flag special considerations that may affect representation, such as whether the client is a senior citizen or veteran.

Legal Server, an online case management platform designed for non-profit and government legal service organizations, exemplifies the benefits of SAI in client intake.14 By incorporating SAI into its online intake module, the automated process has a conversational, human-like approach without requiring staff involvement, resulting in more time those people can devote to tasks that can’t be performed by machines. Coupled with NLI, SAI can become even more adept at interactive conversations.15

With the available technology, the goal should be expanding its use nationwide to provide low-income populations with greater accessibility to legal information and services.

“TXT ME PLZ”

Automated text messaging is certainly not a new concept. We receive automated texts every day from pharmacies, restaurants, and doctors’ offices, but this technology has only recently become part of the discussion regarding access to justice. It is another critical tool in narrowing the access to justice gap. Attorneys working with low-income clients know that the best way to communicate with their clients is via text message, and courts are coming to the same realization with respect to unrepresented parties.

Courts across the country are using automated methods to send parties text reminders of upcoming court dates and other appointments.16 Their stated goals are reducing the number of bench warrants issued (and the resulting expenditures of time and money) and improving case flow by eliminating delays caused by no-shows, while also reducing the collateral effects of bench warrants and defaults on the parties against whom they are issued. Statistics show that the use of automated texting has had the desired effects; failures to appear have decreased in courts using the technology.17

Legal services providers are also working to increase automated texting to follow up with clients and visitors. According to Tripp, MLH launched its Next Steps Text program in July. Visitors who prepared certain forms such as divorce or eviction answers can opt into a series of automated text messages that prompt them to take actions (such as filing or service), remind them of timelines (such as life of a summons), and help MLH learn more about the outcomes for people using its tools. The system can also provide just-intime information or guidance at later steps in the legal process.

Tripp provided an example of how this works. A text may be sent automatically a week after opt-in asking if the litigant filed the complaint drafted on the MLH website. Using branch logic, the system will analyze that person’s response and send another context-appropriate text. If the person responds “yes,” the system may ask questions to ensure the complaint was properly served and/or remind them how long the defendant has to answer. If the answer is “no,” the system may send a link to a page on the MLH website where the litigant can learn more about how to serve the other party.

Automated texting has two primary benefits for access to justice. First, it reaches low-income individuals in a way that makes them more likely to respond. Second, it performs tasks that would otherwise need to be handled by humans, leaving staff more time to do work that machines cannot. Finally, automated texting allows organizations and courts to operate at scale; effectiveness and efficiency lead to better accessibility to legal services and the justice system.


1        Numerous organizations, including the American Bar Association, recommend that hosts of virtual meetings take certain measures to accommodate the needs of the disabled. Several have published best practices, e.g., Virtual Meetings: Accessibility Checklist & Best Practices, ABA Comm on Disability Rights (July 1, 2021) [https://perma.cc/2NBL-7NPK] and Zoom Considerations for Teaching Students with Disabilities, Poorvu Ctr, Yale Univ (2021)

[https://perma.cc/5DWL-GA8C]. All websites cited in this article were accessed September 7, 2021.

2        Michigan’s Digital Divide, The Education Trust – Midwest (2021) [https://perma.cc/7ZQ5-S233] (examining Michigan’s digital divide in the context of education).

3        Black, These document assembly tools will keep your law firm on track, ABA Journal, ABA (June 25, 2019)

 

www.abajournal.com/news/article/these-document-assembly-tools-will-keep-your-firm-on-track> [https://perma. cc/F2RC-AE4X].

4        Barto, What is Branching Logic? Roundtable Learning

[https://perma.cc/NK7P-EWFR]. Many may be familiar with branching logic through its use in do-it-yourself tax preparation software.

5        Michigan Legal Help [https:// perma.cc/4YGA-7CLD].

6        Simms, What is Natural Language Understanding (NLU)? A beginner’s guide, VUX (March 31, 2021) https://vux. world/what-is-natural-language-understanding-nlu-a-beginners-guide/ [https://perma.cc/6NJT-YFK8].

7        Learn more about Spot at The Legal Innovation & Technology Lab’s Spot API, Suffolk Law School (September 30, 2021) [https://perma. cc/KW2W-BY3A].

8        For example, Spot is used in connection with the Massachusetts Legal Resource Finder site funded by the Massachusetts Legal Assistance Corp, [https://perma.cc/KD9N-H9VZ].

9        Natural Language Understanding (NLU), TechTarget.com

[https://perma.cc/ GR6W-2296].

10      See the product website at [https://perma.cc/7YE2-7PKF].

11      Learned Hands, Stanford Legal Design Lab & Suffolk LIT Lab < https://learnedhands.law.stanford.edu/> [https:// perma.cc/DFN6-RFNT].

12      Dickson, What is symbolic artificial intelligence?, TechTalks (November 18, 2019) [https://perma.cc/D5VJ-U57B].

13      What is Branching Logic?

14      See a description of Legal Server’s client intake capabilities on their website at [https://perma.cc/A3LB-UJHF].

15      The Case for Symbolic AI in NLP Models, expert.ai (May 19, 2021) [https://perma.cc/D59Y-VF4R]. 16 Tashea, Text-message reminders are a cheap and effective way to reduce pretrial detention, ABA Journal, ABA (July 17, 2018) [https://perma.cc/XW87-6LVF].

17 Cooke et al, Behavioral Science to Improve Criminal Justice Outcomes, ideas42 and Univ of Chicago Crime Lab (January 2018), available at [https:// perma.cc/D56G-VJBD].