by Kenny Berkowitz
It’s hard to overstate the potential of artificial intelligence. Even at this early stage, three years after the launch of ChatGPT, lawyers are using generative AI to streamline their workflow, summa-rize documents, analyze case records, and draft legal briefs and contracts.
Sometimes the process works well, succeeding quietly, and sometimes it doesn’t, failing spectacularly. Either way, the law’s reliance on precedent makes it an obvious application for AI, and there’s good reason to believe the technology is here to stay. Given enough time, it promises to save labor for law firms and increase access to justice for clients—but as the capabilities of AI increase, so do the possibilities for error.
AI can unknowingly hallucinate, finding cases that were never filed, facts that don’t exist, and outcomes that aren’t grounded in any kind of reality. It can effectively retrieve some court decisions while missing others, and it can fail to understand subtle and not-so-subtle differences in legal reasoning. It can construct legal argu-ments that superficially seem to make sense, even when they don’t, and with all the changes happening so quickly, everywhere at once, academia is struggling to keep pace.
There’s been a strange stasis in legal education on AI, and to me, it seems obvious Cornell Law needs to become a leader in the field.
Jens David Ohlin, Allan R. Tessler Dean
“There’s been a strange stasis in legal education on AI, and to me, it seems obvious Cornell Law needs to become a leader in the field,” says Jens David Ohlin, Allan R. Tessler Dean and Professor of Law, talking about the Law School’s new Center for Law and AI. “Everyone recognizes AI is going to have a transformative impact on almost every aspect of the profession. But most law schools are doing precious little in devoting resources, thinking about the ramifications, or training students to become effective practitioners in this new environment—and I’m not sure why.
“Maybe law schools are waiting for things to shake out before they devote too much attention to it, or maybe they’re paralyzed with uncertainty about the future,” continues Ohlin. “All I know is that there are national conversations that need to be had about the role of artificial intelligence in the legal system, and Cornell is the logical place to start that dialogue. There’s a generation of law students who need to be prepared for this brave new AI world. The time for waiting is over.
Last spring, Ohlin sat down with faculty member Jed Stiglitz, who proposed leveraging the Law School’s existing strengths to found a center focused on law and artificial intelligence. In Stiglitz’s vision, the center would bring together Cornell Law faculty who are researching AI, Cornell Law faculty who are researching with AI, and Cornell faculty from all over campus, especially the College of Arts and Sciences, Bowers College of Computing + Information Science, and the Brooks School of Public Policy. Given a little more time and resources, the Center would expand to include grad research assistants and post-docs, help cover the costs of computing and servers, and begin hosting conferences with experts from around the world.
Then in September, mere weeks later, the Center opened its website and virtual doors.
“That’s light speed in academic terms,” says Stiglitz, who directs the Center, along with working as associate dean for academic affairs and the Richard and Lois Cole Professor of Law. “Cornell already has arguably the best, strongest intersection of law and AI anywhere in the country. The Center is going to build on that, centralizing efforts, fostering collaborations across disciplines, and allowing us to develop that strength.
“It’s not enough for students to be fluent in using these models,” continues Stiglitz. “We need to provide an avenue for them to be the best morally based, technologically enabled lawyers of the future. They need to be sufficiently cautious that they can use models responsibly as the technology grows. They need to under-stand the capacity, the potential of AI, while also recognizing its limits. We can offload some tasks to bots, which are good at discovery, at extracting information from long documents, and at summarizing data. But we can’t offload the moral dimension, the human dimension, which is what we’re here to teach.”
Now in its second semester, the Center currently comprises ten Law School faculty—Yun-chien Chang, Michael Dorf, Jessica Eaglin, James Grimmelmann, Sarah Kreps, Karen Levy, Kim Nayyer, Frank Pasquale, Stiglitz, and Kristin Underhill—along with affili-ates from Brooks, Computer Science, Duffield Engineering, Government, Information Science, Linguistics, and Statistics and Data Science.
Working with Ohlin and Thorsten Joachims, the University’s new vice provost for AI strategy, Stiglitz has outlined four main focus areas for the Center: AI and Legal Education, AI and Legal Practice, AI and Scholarship, and Regulating AI. With the first prong, researchers will explore ways for legal pedagogy to integrate computational thinking and AI literacy into the classroom; with the second, they’ll investigate present and future changes to attorneys’ roles, the structure of legal work, and the gap in access to justice.
AI and Scholarship probes questions about how AI can discover patterns in legal texts, generate and test legal theories, and expand the empirical study of law. Regulating AI examines how domestic and international legal frameworks can be designed to responsibly govern AI, scrutinizing design, institutional capacity, democratic accountability, and the evolving role of law in shaping the deploy-ment of AI technologies.
“We’re living in a time where existing practices are impacted by artificial intelligence,” says Joachims, the Jacob Gould Schurman
Professor in the Department of Computer Science and the Department of Information Science. “The ability to use AI as a tool is coming within reach of more people than ever before. We cannot ignore AI, and so I encourage the disciplines to engage in construc-tive dialogue to understand where AI can provide benefits, how to design effective AI systems, how to use them responsibly—and where to not use AI.
“Cross-disciplinary collaborations like ones in the Center are an essential part of this dialogue,” continues Joachims. “The institu-tions that effectively respond to these new opportunities and challenges in the most thoughtful, immediate ways are the ones that are going to establish leadership in the field.”
Starting with a simple understanding—that law schools should be guiding the transformations of AI, not reacting to them—Stiglitz envisions the Center as a tool to probe new technologies, support scholarship, and ensure the law serves as a source of fairness and democratic oversight. He sees lawyers and legal professionals as having a dual role: on the one hand, they need to help design ways for legal institutions to utilize AI, and on the other, they need to “serve as the conscience of that transformation,” finding ways to examine and overcome the biases that can be amplified by the algorithms.
It’s not enough for lawyers and legal scholars to use generative AI without understanding how it works, and it’s not enough for tech-nologists to work in isolation, without the doctrinal, historical, and empirical insight of legal academia. It’s not enough for students to apply the technology to completing their next assignment; they need to combine technical proficiency with a complex, sophisticated legal perspective to develop an understanding of what AI can and should do.
“As a society, we finally have the ability to scale human-like intelligence in ways that are incredibly exciting and incredibly disruptive,” says Stiglitz, whose teaching and research focuses on administrative law. “Now we need to learn how to adapt, how to bring legal reasoning and technological expertise into a conversa-tion, and how to use the strengths that are already here to create an intellectual hub for change.”
That was all Ohlin needed to hear.
“There’s universal agreement among lawyers that AI is going to radically transform the practice of law,” says Ohlin. “I’ve never encountered a topic where there was such a strong consensus—and still so many divergent opinions about how things are going to play out. There’s almost universal disagreement about how the practice of law will be transformed. What will it look like? What will the impacts be? What challenges will we face?
“This isn’t something that’s just affecting the law,” continues Ohlin. “It’s a broad societal change, and there’s a university-wide commitment to bring people together. Not only do we have lawyers in the room where those conversations are happening, we’re having conversations here at the Law School, bringing in key partners from other divisions and building the partnerships that are going to be an essential part of our success. There’s a lot of thinking that needs to be done, and Cornell Law is committed to being on the leading edge.”
Building the Future
In its first few months, the Center for Law and AI has found its footing with a series of interdisciplinary collaborations already started and others still in the planning stages. Next comes the hard work of building an infrastructure to keep up with the pace of technology.
“Nothing comes close to approximating the dynamics we’re seeing with artificial intelligence,” says Director Jed Stiglitz. “This field is moving so fast and so dramatically, it’s unlike anything else. In that sense, it’s not possible to plan ten, fifteen, twenty years down the line. The immediate goal is for our research to grow with the changes, to make sure we’re using the most up-to-date models in the most responsible ways, and to effectively communicate those lessons to our students and the world.”
Stiglitz’s wish list begins with funds for a series of conferences, workshops, and colloquia, including events proposed by faculty Yun-chien Chang, Jessica Eaglin, James Grimmelmann, Kim Nayyer, and Frank Pasquale. Next are a set of executive education courses offered to judges, legal practitioners and scholars, librarians and information specialists, professional education managers, and corporate legal specialists. Funding for research assistants and postdocs could support projects to analyze the language used in capital trials, evaluate the accuracy of courtroom audio files, and compare constitutions from around the world. Funds for teaching faculty could be used to support the creation of a clinic.
“Cornell Law is filled with deep, ambitious thinkers who are focused on the nature of the law and on the legal system,” says Ohlin. “They’re not shutting themselves off from the real world of legal practice, and with the growth of artificial intelligence, they’re re-conceiving the ways we interact with the law. They’re making sure our students are being trained for a world of legal practice dominated by artificial intelligence, and they’ve made the Center a top priority for all of us.” ■
