Prior to FutureLaw, the inaugural UN AI for Good conference tackled the power of AI to promote sustainable, ethical, and responsible law and legal systems, building on the United Nations’ AI for Good initiative launched in 2017. As in past years, a two-day hackathon and technical bootcamp attracted students from across Stanford University, along with hundreds of legal-tech industry members, to innovate at a feverish pace with large language models (LLMs). A CodeX Affiliate Showcase offered presentations and short demonstratives, workshops, and discussion groups on all manner of innovations over the last two decades, from AI adjudication and agents to computational antitrust.

“Emotions were high this year,” said Megan Ma, a CodeX fellow and associate director of CodeX and the Law, Science, and Technology Program, as well as an organizer of and speaker at FutureLaw. “There was a very strong sense of community and deep introspection this year as we took stock of how industry leaders, students, and faculty have come together in this hub of innovation.”

Larry Kramer, dean of Stanford Law at the founding of CodeX and now vice-chancellor at the London School of Economics, recalled that the synergies between Stanford Law and the Computer Science Department were immediately obvious when CodeX was established 20 years ago. CodeX would not just be for research, Kramer said in the anniversary video, but a place where “people could bring ideas for changes in technology – things that could be done to improve legal services. And we would help nurture these ideas and take them out into the world. No other law school had really done this, but it fit so nicely with the ethos of Stanford and Silicon Valley. That’s the Stanford spirit that CodeX has brought forward: an interdisciplinarity way to think outside the box.”

Innovation in action

The annual CodeX hackathon, co-led by Ma, CodeX fellow Jay Mandal, and visiting scholar Pierre-Loic Doulcet, kicked off FutureLaw Week. CodeX launched its first LLM x Law hackathon in 2023, marking the first time that a hackathon was held, anywhere in the world, at the intersection of LLMs and the law. This year, approximately 400 “hackers” participated, including close to 80 Stanford University students representing law, engineering, computer science, linguistics, and business, among other disciplines. Over two days, including a day of hands-on training (aka “the Bootcamp”), interdisciplinary teams competed to come up with new ideas for streamlining, automating, and augmenting legal practice.

The hackathon participants’ work was judged by representatives from law firms and tech companies. Summer Shaw, JD/MBA ’25, and Eduardo de los Heros, LLM ’25, were on the team that took home the Best Overall prize for an AI tool they conceived and dubbed Gov Match. The AI tool helps small- and mid-sized businesses find and respond to government requests for proposals (RFPs).

“U.S. federal, state, and local government spends over $2 trillion each year in contracts, of which nearly 25 percent is specifically earmarked for small- to medium-sized businesses; however, this spending remains frustratingly elusive for many small businesses, which struggle to navigate government RFPs, which can have hundreds of pages of requirements,” explained Shaw. “Gov Match helps these companies win more government contracts by utilizing AI-powered semantic search to pinpoint the contracts that are the best fit for their company and also leveraging a multi-agent AI model to generate a draft response to a given RFP, reducing the time required to prepare a response from days or weeks to just minutes.”

De los Heros, Shaw’s teammate, said the hackathon experience was “exciting and exhausting.” Once the team landed on their concept, “the rollercoaster truly began,” he said. “At times, I felt incredibly energized and motivated, at others, completely drained. That emotional ride, combined with the challenge of stepping far outside my comfort zone in such a short period, made the experience both intense and incredibly rewarding.”

Adrian Mak and Soraya Pradhan, both LLM ’25, along with their Arb Litt Up team, received awards for their legal argument matching system, which uses AI to help legal professionals by automatically finding semantically related arguments between plaintiff and defendant documents in cases like arbitrations. “The costs of matching arguments from thousands of pages on legal briefs, assessing argument strength, and repeatedly re-prompting AI models for precision can be immense,” Mak said. “Our approach builds expert evaluation directly into the loop, creating a high-touch system that learns from feedback what works best. This significantly cuts down the costly back-and-forth, giving legal professionals more time to focus on the key results and strategic decisions.”

For the third time, the hackathon included a “VC Track” that allowed past participants to build on their work from prior hackathons and get their ideas in front of venture capitalists.  

“CodeX has always been focused on practical applications of AI and introducing ideas to the industry, so this track at the hackathon is a natural next step to support our repeat participants and provide a form of mentorship for them in this space,” Ma said. “There’s so much forward momentum on these great ideas.”