AI-powered NLP enhances legal aid at Justice Connect

Australian non-profit Justice Connect, dedicated to providing legal assistance to disadvantaged individuals and advocating for social justice, has integrated natural language processing (NLP) technology to enhance its operations and extend its reach.

NLP, a branch of artificial intelligence (AI) focused on enabling computers to understand human language, has been pivotal in increasing the organization's efficiency and helping more people access crucial legal support.

Extensive network and mission

Justice Connect collaborates with over 10,000 pro bono lawyers, aiding thousands of individuals annually with issues ranging from housing and employment disputes to elder abuse and family violence. The organization also offers practical legal resources online and advocates for systemic changes to unjust laws.

However, categorizing and assessing each unique case to provide the most effective assistance was historically a time-consuming manual task. This changed with the involvement of Professor Timothy Baldwin, an expert in large language models (LLMs) and a Melbourne Laureate Professor at the University of Melbourne, who saw the potential for an NLP-based solution to streamline these efforts.

Recognizing inefficiencies

Tom O'Doherty, Head of Innovation at Justice Connect, recognized inefficiencies in the organization’s digital intake tool, leading to unnecessary resource expenditure and delays in service delivery. "Our mission is to provide essential legal advice to Australia's most vulnerable populations," he explained, emphasizing the importance of maintaining efficient and effective systems.

Professor Baldwin, also a Professor of NLP at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), contributed his expertise in high-impact LLM products to this initiative. They developed and integrated NLP models with Justice Connect, yielding significant positive outcomes.

Training the model

Justice Connect collected anonymized data from their intake tool to train the NLP-based solution. They created a 'tagging game' where legal partners tagged and categorized legal requests, refining the data to teach the model to triage casework. This supervised learning process resulted in the tool effectively understanding various legal queries and directing them to appropriate solutions.

The impact of this integration over the past year has been substantial, with a 50% reduction in incomplete entries and a 10% increase in service delivery efficiency. This improvement has allowed Justice Connect to assist people more quickly and accurately, with positive feedback from paralegals noting the ease of dealing with cases referred by the AI tool.

Future expansion

Encouraged by these results, Justice Connect is expanding and enhancing the system's functionality to further refine user segmentation and self-help legal advice available on its website.

It also envisions developing an open API for other non-profits in Australia to adopt similar technologies, fostering broader social justice advancements.

Interested in expert templates to help move your career forward?