At VibeODR, an AAA-sponsored hackathon, seven teams had six hours to explore where AI could help resolve real-world legal problems, turning ideas into prototypes that could be tested, questioned, and refined.
For roughly 20 years, Colin Rule, CEO of ODR.com, gave students in his online dispute resolution class the same final assignment: design an ODR platform and explain it in a 10-page paper. The last time he taught the course at Pepperdine, three students went further. “Here’s my paper, and here’s the working app,” Rule recalled them saying.
Brian Abrams, managing partner at B Ventures Group and VibeODR judge, has judged hackathons for 15 years. He noted a sharp break in what teams can now accomplish: “Post-AI, even non-technical teams can create in hours what used to take development teams weeks or months.”
An idea about how a dispute-resolution process might work could now become something people could actually try.
VibeODR put that shift to the test. Held June 13 at Suffolk University Law School in Boston and sponsored by the AAA-ICDR Institute, ODR.com and others, the one-day hackathon brought together participants with widely varying levels of technical experience. Some had never tried this kind of coding; others were experienced technologists. Dispute-resolution experts were available throughout the day as teams formed around problems and began building. By afternoon, the teams had produced very different ideas about where technology might be useful in a dispute.
Where AI Fits Into a Dispute
One team of recent computer science graduates built Medias, a prototype for landlord-tenant disputes. They focused on information and power asymmetries: a tenant may have a lease without understanding its provisions, face a language barrier, or lack the resources available to the other party.
Medias used AI to organize accounts of events, analyze lease information, identify areas of disagreement, and prepare information for a mediator. The team drew a clear boundary around the technology’s role: AI could summarize and supply information; the human mediator would make the judgment call as to how to use that information.
Another team built Amparo.Health to support health-related claims in Colombia. The presenters described people who may formally have a path to represent themselves but struggle to discern what information matters, which documents to provide, or what happens next in the legal process. Their prototype explored how technology could help users navigate those practical steps.
A family-law team worked on recurring post-divorce disagreements, using a child-related expense as its example. Its prototype organized documentation and communication between parents and provided a route to a human mediator when they could not resolve the issue themselves.
Several teams made a similar design choice: putting AI around a decision rather than asking it to make the decision.
Case Compass, for example, explored using AI to organize court case information and assess whether certain matters might be appropriate for mediation; the clerk retained authority to choose the path forward. Medias reserved the judgment call as to how to use that information for its mediator.
That is a useful distinction in dispute resolution, where discussion about AI often jumps quickly to whether technology could replace lawyers, mediators, arbitrators, or judges. The VibeODR teams spent much of their time on narrower problems inside existing processes: helping someone understand information, assembling relevant material, or getting a dispute into an appropriate process.
Making Access Practical
A legal process may technically be accessible but prove difficult to navigate. Amparo.Health began with that problem. The family-law team approached it from another perspective. One of its law student members, drawing on previous work with self-represented parties navigating legal problems, recalled clients who felt they had “no control over their court case,” with decisions seemingly resting with lawyers, mediators, or judges they did not know or trust. Medias addressed a related friction: having a lease is not the same as understanding one’s rights when a dispute arises.
None of these prototypes established that AI would solve those problems. They did make proposed responses specific enough to examine.
From Idea to Cross-Examination
That examination began before the day was over. Case Compass initially asked parties a yes-or-no question about their willingness to mediate. A judge suggested that willingness might be better expressed on a scale. The family-law team was asked what would make highly contentious parties use its proposed process. Medias fielded questions about what incentives would induce parties to participate.
The mix of expertise in the room shaped those conversations. One team working on safeguards involving minors and AI included an arbitrator whose background included investigative, prosecutorial, and legal work. He readily acknowledged that the technical side of AI was “definitely not my forte.” His teammates brought technical capabilities; he contributed experience identifying legal and safety concerns.
On the family-law team, the law students worked alongside John Lande, a retired law professor who has studied dispute resolution since the 1980s. He connected their prototype to parenting coordination and began considering where the approach might help—and where its limits might lie.
Dispute-resolution technology has to operate within legal processes while accounting for how people actually behave inside them. A prototype gives technologists, lawyers, dispute-resolution practitioners, and others something specific to react to—and something concrete to challenge.
What Emerged in Six Hours
Seven teams produced prototypes spanning landlord-tenant matters, family disputes, health care, online safety for minors, neighborhood conflicts, and other legal problems. Hackathon judge and ATJustice CEO David Cohen noted that choosing among them was difficult because “every single entry was a legitimate contender.”
The judges ultimately selected Mending Fences, a prototype focused on everyday neighborhood disputes, as the winner. Abrams also saw potential beyond the competition. Ideas that once might have ended as presentations could leave the room as working concepts that teams could continue testing—and, in some cases, develop into startups.
The seven prototypes did not establish that the proposed systems would improve outcomes, reduce costs, or expand access in practice. Some features were unfinished; some assumptions were challenged before the presentations had even ended. Teams had to make early choices that a proposal might leave unresolved: what information a user would provide, what AI would do with it, when a human would step in, and why another party would participate. In prototype form, examination of a tool’s effectiveness bears fruit much faster and accelerates the point at which an idea had to withstand scrutiny—the ultimate measure of hackathon success.
About the AAA-ICDR Institute™
The AAA-ICDR Institute™ is the "think and do tank" of the American Arbitration Association (AAA). We shape the future of alternative dispute resolution (ADR) through applied research, partnerships, and strategic initiatives. The AAA-ICDR Institute generates pioneering insights and develops practical, evidence-informed solutions involving AI, access to justice, and court-connected ADR that advance fair, efficient, and accessible systems for resolving disputes.