RelativityOne Is Now  Relativity aiR logo
We rebranded our platform to reflect the growing role of AI in legal work. Our blog reflects a years-long library of insights, so don't be surprised if you spot our former name, “RelativityOne,” here and there. Learn more
Relativity Home logo

Your single source for new insights on AI, legal data intelligence, and the humans holding the reins.

A Closer Look at AI in Legal Education and the Law

Clair Atkins
A Closer Look at AI in Legal Education and the Law Icon

When Relativity Academic launched its first webinar series this spring, the goal was straightforward: help professors and academic leaders understand not just that AI belongs in legal education, but how they can realistically bring it into the classroom.

Our first session, From Interest to Implementation: AI in the Law School Curriculum, was a frank accounting of where legal education can fall short on AI, what practicing lawyers need from new graduates, and what it looks like when a law school commits to building AI fluency in the formative years of a budding legal career.

Our second webinar, Professor Perspectives: AI and the Law, shared that theme but shifted the lens just a bit: how can legal education thoughtfully incorporate AI while still fostering the critical thinking and analytical skills the profession depends on? The goal was to leave our audience with practical strategies, ethical considerations, and concrete next steps that could benefit their students right away.

Before diving into the highlights of each session, we want to acknowledge the concern we hear from both faculty and students. Will there be a role for me when I graduate? How do I bring AI to my students in a way that helps them prepare for the future?

Whether you watch the webinars, read this summary, or both, our hope is that you come away knowing that humans power AI, and that using AI well means bringing your own perspective and creativity to it. Learn as many tools as you can and look for how each one can help you work faster or turn an existing role into something new. That’s how you lay the foundation for a successful, fulfilling career in the legal profession.

From Interest to Implementation: AI in the Law School Curriculum

Joining Relativity's Aron Ahmadia, vice president of applied science, were two guests with firsthand perspectives on both sides of the issue: Oliver Roberts, CEO and founder of Wickard.ai, practicing attorney and adjunct professor at Washington University in St. Louis; and Stefanie Lindquist, Nickerson Dean and professor of law at Washington University in St. Louis School of Law.

Roberts was direct: the expectation in litigation is already here. Summer associates and recent graduates are expected to know the basic functions of research tools, and the generative AI features now built into both. The expectation goes beyond simply using the tools, though – new professionals have to be realistic about their risks and limitations as well.

“There's a common misconception that if we're using a legal tech-specific product, there's no hallucination because they're hyperlinked, grounded in real cases,” Roberts said. “But the propositions or quotes attached to those real citations could be totally fabricated.”

Aron added a striking data point from Relativity's own community: “As of this year, AI systems in the Relativity platform make more decisions on documents than humans do.” He described telling first-year associates encountering aiR Review: “I'm giving you a tool that allows you to review a million documents over a weekend. That used to be a six-month project that would cost over $1,000,000. You should feel a little scared.”

Lindquist pointed to a gap many schools share: project management. “One thing we don't do particularly well in law schools is teach project management. We don't teach an entrepreneurial or research and development mindset either,” she said. Roberts agreed, recalling his own start in practice: “I couldn't even tell you what goes into a motion to dismiss.”

Asked which AI skills won't be obsolete in two years, Roberts named three: a foundational understanding of how generative AI works, grounded knowledge of legal ethics, and an adaptive mindset. Lindquist added the Socratic method: “Law students need to be able to think like lawyers before they delegate that thinking to a model. If law students use AI too much, too soon, they may lose that capacity to be excellent.” AI, she noted, can draft a strong motion to dismiss, “but it may miss creative opportunities” that come from a well-trained lawyer who knows how to make facts and law fit together in new ways.

Lindquist built the school's AI Collaborative from her first weeks on the job, bringing on Roberts to train students, faculty, and local practitioners, and later hiring a large language model engineer. Roberts's course runs through four stages: foundational concepts, practical use of general-purpose tools, ethical responsibilities, and hands-on exercises built around realistic fact patterns – plus, most recently, “vibe coding” competitions where students build legal-tech prototypes from scratch.

For schools starting from zero, Roberts pointed to the AI already embedded in research platforms as the lowest-hanging fruit, and warned that vendor partnerships only work with “an academic point of contact and a program.” Lindquist's frame for skeptical faculty: “The practice of law is driving AI adoption, not law schools … the profession is going to keep moving, regardless of what any individual professor decides to teach.”

Professor Perspectives: AI and the Law

Our second panel brought together Dean McPeak, associate dean for Academic Affairs and Program Innovation at Gonzaga University School of Law; Judge Rodriguez, U.S. District Judge for the Western District of Texas and adjunct professor at St. Mary's University School of Law; and David Horrigan, Relativity's discovery counsel and legal education director and an adjunct professor at Duquesne University.

Dean McPeak named the contradiction many schools sit in: banning AI while still expecting students to graduate into firms that use it heavily. Her answer to this challenge? Scaffolding: starting with the fundamentals of thinking like a lawyer, then introducing AI strategically, course by course.

Gonzaga now requires a one-credit course, AI Legal Research Methods, that teaches students to fold AI appropriately into each step of legal research. Some Gonzaga professors ban all technology in class; others build entire courses around AI from day one. “From an academic freedom standpoint, I support it all,” she said.

Judge Rodriguez described surveying large law firms about their expectations of new graduates, since most students won't land at firms with training budgets to match. “Everybody's not going to be in big law,” he said. “If they're not going to get that training, then we have to give it to them.”

David Horrigan walked through Duquesne's three-tier AI policy, which lets professors choose an open, moderate, or restrictive statement for their own classroom. But Judge Rodriguez raised a hard truth: policies only matter if they can be enforced, and detection tools aren't reliable enough yet. A Northwestern study found deep-fake detection tools failed six or seven times out of ten. Even journalist Bob Ambrogi, Horrigan noted, found his own, entirely AI-free writing flagged as AI-generated when he ran it through a detector himself.

Two real cases underscored the stakes. In Ogletree v. Cleveland State, a student successfully argued that AI-powered exam-monitoring software scanning his bedroom during a remote test violated his Fourth Amendment rights. And in a more recent Yale matter, a student was accused of academic dishonesty based on an AI-detection tool that, as in Ambrogi's case, produced a false positive. Horrigan's takeaway was that these tools “are as good as the people who use them … but they can be dangerous if used the wrong way.”

Rather than chasing the “right” tool, Dean McPeak argued for a task-based approach that answers “what are you actually trying to do, and what do you have access to?” questions. She has students co-author a living AI policy for her own classes, sorting real use cases into acceptable, borderline, and unacceptable columns. A document she then finalizes as the instructor. It surfaces use cases she hadn't anticipated, like quiz generation and AI-built study podcasts, while reinforcing the idea that tools are okay, but shortcuts aren't, because AI can't substitute for the substantive work of learning.

Both McPeak and Judge Rodriguez flagged access and equity by calling out the fact that many students only have free or FERPA-approved tools like Microsoft Copilot, while classmates headed for big firms may get hands-on with paid, specialized platforms. “We're not providing access the way we do Lexis and Westlaw,” McPeak said, “so I don't want to say you have to use certain tools if I'm not providing them.”

If AI makes take-home work hard to verify, the panel's answer is to change how students are assessed, not only what they're allowed to use. Expect more oral exams, proctored closed-book tests, and scaffolded assignments that force visible work along the way. McPeak also floated swapping graded “class participation,” which she called too subjective, for graded class presentations. This manifests as an exercise where AI simply can't help once a student is standing at the podium. Her closing advice to faculty: “Don't assume your students want a shortcut. If you emphasize the learning and personal growth, they'll recognize that AI deprives them of that.”

Asked what AI vendors could do better, Judge Rodriguez asked for more transparency on plain-language questions, like whether prompts train the model and how to control or delete data, plus less of the flattering, overconfident tone that can make AI answers feel more authoritative than they are.

Closing the session, each panelist named one thing they wish students understood beyond “verify your citations.”

For Horrigan, it's that AI is not an “easy button. Talented engineers built it, but it isn't foolproof.”

For Judge Rodriguez, it's a basic grasp of how generative AI really works. When a court asks a lawyer to explain a model's “black box” reasoning, he said, “I'm sorry, it's a black box” isn't an acceptable answer.

And for Dean McPeak, the answer was simple: students need to know that they are human, and AI is not, because “we will never substitute the value that human lawyers bring to people in need.”

Key Takeaways

Across both webinars, a consistent thread emerged: AI is reshaping what law schools must teach and how firms operate, but the profession's most durable skills remain distinctly human. Those skills include ethical judgment, creative legal reasoning, and the ability to sit with a client and understand what they need.

Schools that scaffold AI into the curriculum thoughtfully, rather than banning or embracing it wholesale, seem best positioned to graduate students who can use these tools well and know their limits.

Graphics for this article were created by Caroline Patterson.

Learn More about Relativity's Academic Programs

Clair Atkins is a lifelong learner and educator with over two decades of experience in the e-discovery and legal technology industries. Since joining Relativity in 2015, Clair has served as a pre-sales engineer and senior manager of customer enablement, where she helped customers transition to RelativityOne. Today, Clair is part of the Relativity Academic Program and works closely with professors, students, scholars, and mentors to guide them in their use of and involvement with Relativity.

The latest insights, trends, and spotlights – directly to your inbox.

The Relativity Blog covers the latest in legal tech and compliance, professional development topics, and spotlights on the many bright minds in our space. Subscribe today to learn something new, stay ahead of emerging tech, and up-level your career.

Interested in being one of our authors? Learn more about how to contribute to The Relativity Blog.