When Arizona State University’s Sandra Day O’Connor College of Law appointed Sean Harrington to lead its newly established AI and Legal Tech Studio in late 2025, the move signaled more than a faculty hire. It marked a step toward reimagining how lawyers will be trained and how legal services will be delivered in an era where AI can complete in minutes work that once took hours.
ASU Law, ranked No. 1 in innovation for the 11th consecutive year, has become a testing ground for what legal education must become as large language models transform the practice of law. The school was the first in the U.S. to explicitly allow generative AI in its admissions process and the first to offer an AI emphasis across multiple degree programs.
But according to Harrington, who previously served as director of technology innovation at the University of Oklahoma College of Law, the real transformation isn’t in how law schools talk about AI—it’s in whether students actually know how to use it when they graduate.
Law, unlike fields such as robotics or manufacturing, is built entirely on words. That simple fact is why Harrington believes AI will transform legal practice faster and more completely than almost any other profession.
Legal text already exists in “giant databases,” he said, easily chunked into tokens for training large language models. AI quickly became proficient at tasks lawyers used to spend hours on — document review, organizing depositions, identifying relevant case law.
This isn’t theoretical. Students at ASU are already building AI-powered legal tools. And one of Harrington’s former students created a platform to help people prepare for law school in about six months, complete with cold call simulations. Another recent graduate launched a meme-based bar study platform.
Training lawyers to build, not just brief
ASU Law offers about 15 AI courses under a certificate program, including specialized classes in AI ethics, legal operations, intellectual property protection for emerging technologies, and blockchain. All first-year law students now receive mandatory training on AI tools as part of their coursework.
The school also launched a Master of Legal Studies in AI and Law, a 30-credit program designed for professionals seeking expertise in law and technology without requiring an LSAT, GRE or GMAT for admission.
But Harrington argues that law schools’ biggest gap isn’t teaching ethics or theory — it’s getting students “actually practically using it” and getting “reps with the technology.”
Current assessment methods are “all broken now because of AI,” he said, requiring faculty to rethink how they teach and test. That’s a difficult ask given academic freedom and tenure, but one that’s becoming unavoidable. Students entering law school now often have three years of AI exposure from undergraduate programs and expect to be taught how to use it, not just warned about it.
Starting this spring, ASU Law will launch the first part-time, online JD program in the U.S. focused on public service. The program uses a proprietary AI platform, built with Ph.D. graduates from ASU’s engineering school, to monitor rigor and triage the learning experience in real time.
The platform provides 24/7 access to AI chatbots trained on professors’ knowledge, simulating a more in-person experience. The goal is to expand access to legal education for people who couldn’t otherwise attend in-person programs while maintaining academic rigor.
Harrington emphasized the program’s focus on addressing Arizona’s “rural access to justice issue,” particularly given the state’s extensive reservation land. The program aims to train attorneys who can serve communities where traditional law school access doesn’t exist.
The billable hour is dead
Perhaps Harrington’s most emphatic prediction is this: The billable hour is “dead” in the age of AI.
When AI can draft contracts, review documents, and research case law in minutes rather than hours, charging clients by time makes no sense. The legal industry is shifting toward value-based billing and subscription models, particularly for in-house counsel.
Harrington pointed to startup Alt Fee, the winner of the ABA Tech Show’s startup competition, which calculates lump-sum fees based on historical firm data rather than hours logged. He anticipates “huge winners” who scale and turn specific legal services into products — companies offering joint summary dissolution for $99, for example.
Law practices will become more modular, with AI tools plugging into existing workflows, he said. Harrington pointed to cloud legal software firm Clio’s acquisition of vLex, a global legal research platform, as an example. The consolidation creates what he called an attractive “operating system” where attorneys “can draft out of it, you could bill out of it. You can text the client out of it. You can do your research out of it.”
While AI will become capable of performing nearly all legal tasks, Harrington said regulation driven by public comfort and accountability will limit AI’s role in positions such as arguing in court or being the attorney responsible for filing petitions.
But entire categories of jobs will cease to exist. Human document review, for instance, is already disappearing. Harrington disagreed with the common saying that attorneys using AI will beat those who don’t, noting that in some areas, there simply won’t be human jobs left to compete for.
“There’s going to be some guy who’s going to use OpenClaw to essentially run a firm.”
New roles will emerge — legal operations specialists, DevOps professionals who bridge law and technology, and positions connecting legal teams with senior leadership. These roles will be less about pushing papers and more about being connective tissue between law and technology, he said.
While large firms can afford dedicated AI teams and training programs, small and midsize practices face a steeper climb. AI adoption is “so pay-to-play,” Harrington said, with high costs for tools and compute creating barriers.
But the bigger issue is time. AI demands resources and time for researching and evaluating workflows, he said — something lawyers have traditionally not allocated to technology.
Harrington’s solution: Firms should designate an “AI champion” with protected time to research tools, test workflows, and determine what actually adds value. Even large firms are recognizing this need. Harrington has been conducting training sessions at AmLaw 100 firms where old methods of training associates are proving insufficient for AI-enabled practice.
An OpenClaw moment for law
When asked about his 2026 predictions, Harrington forecast an “OpenClaw moment” for law — a point where someone uses an autonomous AI assistant to run an entire law firm. Unlike traditional AI tools that require human direction, OpenClaw can independently execute tasks across multiple systems.
“There’s going to be some guy who’s going to use OpenClaw to essentially run a firm,” Harrington said. “He’ll plug it into enough APIs. He’ll be able to answer his phone, do his billing and do all that stuff.”
Harrington, who has set up his own OpenClaw system, described its autonomous capabilities: “I can text it. Go do research. Go read my emails. Go check Melissa’s website. Go send Melissa $20 from my bank account. It can do all that just from a text.”
The system can even conduct legal research on Westlaw and perform other tasks without explicit permission for each action. While he acknowledges the implementation would be “clunky here and there,” Harrington believes this demonstration will force the profession to confront what truly autonomous legal practice looks like.
The transformation won’t happen uniformly. Adoption of AI appears to be more “curiosity based” than age-based, Harrington noted. About a third of current law students remain “allergic to this stuff” due to concerns about sustainability, intellectual property, and ethics.
But for law schools like ASU that are betting on AI integration, the stakes are clear. Students are entering a legal market where AI literacy isn’t optional — it’s foundational. The question is whether law schools will prepare graduates to lead that transformation or force them to catch up after they’ve already started practicing.
For Harrington and his students building AI tools at ASU’s Legal Tech Studio, the future of legal practice isn’t a distant possibility — it’s what they’re coding today.
