State funders and tech partners outline steps to build AI capacity in legal aid

State-level funders and technology partners at a conference session urged legal aid organizations to prioritize basic cybersecurity and deliberate training as they adopt AI tools.

"AI is not a someday issue anymore," Lisa Colpoise, senior consultant to the AI initiative for the Lawyers Trust Fund, said as she opened the session, framing the program as one focused on practical programming to build confidence and competence among legal aid professionals. Speakers from California, Illinois, Michigan and New York described initiatives that pair funding with training, peer networks and technical assistance.

MJ Joyce Smith, a representative of the State Bar of California, described the Legal Aid Justice Technology Collaborative (LAJTC) and its three priority areas: supporting baseline technology adoption, fostering leadership and community among grantees, and creating tools to support responsible AI use. "We are exploring a statewide cybersecurity program modeled off of the Legal Services Corporation's KnowBe4 program," Smith said, and added the state is considering rule changes that could require additional cybersecurity standards for grantees.

In Illinois, Colpoise said the Lawyers Trust Fund (the state's IOLTA program) launched an AI initiative in 2024 that combined discovery interviews, webinars, peer learning groups and strategic opportunity grants to help 57 grantees begin experimenting. She described focused pilots including AI phone agents for intake and automation of document workflows. The Lawyers Trust Fund also ran a survey that Colpoise said drew responses from about 25% of the state's legal aid professionals, roughly 400 people, showing a notable increase in familiarity and use of AI tools within a year.

Bridget Caron, a professor at the University of Michigan Law School, urged leveraging academic partnerships and student teams to incubate prototypes and serve as supervised humans-in-the-loop so overburdened staff can continue direct service work while tools are iterated.

Ellen Samuel, chief operating officer at Just Tech, outlined New York's cybersecurity-and-technology advancement project (CITAP) and its AI University cohort. She said 72 organizations participated in the AI University hands-on labs, where participants received paid ChatGPT licenses to test use cases. Samuel described a product the firm developed, "JT AI," to build searchable knowledge hubs and deploy AI agents that answer client queries (not legal advice).

Across speakers, two consistent recommendations emerged: ensure baseline security and data hygiene before deploying client-facing AI, and use cohort or shared-license models to reduce costs and spread training. Presenters also recommended peer-to-peer learning and champion networks so practitioners learn from colleagues with similar caseloads.

During Q&A, participants raised access concerns—whether chatbots and web tools would be usable by clients who lack internet access or devices—and funding strategies, including shared purchasing of licenses and funder-led training. One attendee noted a funder in Florida plans to provide a Westlaw "co-counsel" AI product statewide as a funded resource.

The session closed with presenters urging blended approaches—local cohorts supported by national resources—and encouraging funders to invest in security, training and shared infrastructure to broaden safe AI adoption in civil legal aid.

The session included no formal votes or policy adoptions; presenters said follow-up resources and slide decks would be available in the conference app.