Northwestern IT Brings Generative AI “Clients” to Law School Classroom
Posted Date: February 25, 2025
Text messaging is becoming increasingly important in practice, and we were excited that our students could practice this skill as they do work for their simulated clients, Andy and Dr. Gold.”
Esther Barron
Director of the Donald Pritzker Entrepreneurship Law Center
The rapid adoption of generative AI in higher education has brought challenges for professors, but it also can bring about solutions and augment classroom learning. Entrepreneurship law students at the Northwestern Pritzker School of Law got first-hand experience with this when the Law School teamed up with Northwestern IT to bring a custom generative AI chat simulation tool to the classroom to support client role-playing and enhance student interviewing experiences.
About the Project
The Northwestern Pritzker School of Law approached Northwestern IT to explore ways to incorporate emerging technologies into the classroom. Northwestern IT Media and Technology Innovation (M&TI) hosted a technology showcase with the Law School to demonstrate the emerging technology capabilities available like virtual reality, augmented reality, and generative AI.
In the initial conversations, the idea of a virtual reality (VR) simulation was tossed around but after more discussion, it was decided that VR wasn’t going to be the ideal user experience in the classroom. Thinking about what could be done with generative AI, M&TI collaborated with their Northwestern IT colleagues in Research Computing and Data Services (RCDS) to hear more about how that team has been using AI with researchers and what could be done in the classroom.
Project Challenge
Getting accurate information from generative AI models can be a challenge, so there was some justified skepticism about bringing generative AI into the classroom. Knowing that there are other use cases for generative AI that don’t require it to be an accurate information source, Northwestern IT identified an opportunity to use generative AI as part of an entrepreneurship law course.
In the entrepreneurship law course, Professors Stephen Reed and Esther Barron role-play as Andrew (Andy) Orlando and Dr. Olivia Gold, two friends who want to create a business together. The students, in turn, act as their lawyers, helping them set up their business. This classroom exercise has been successful because it lets students ask questions face-to-face with their clients and gives them a chance to read their clients’ body language. Students can also benefit from brainstorming with each other and see their peers’ line of questioning and framing.
The in-class exercise serves as an effective form of practice for the students, but it also has some key limitations. There isn’t enough time in class for all students to ask all of their questions. Most students only get to ask one or two questions in the classroom, which limits the practice each student gets. Additionally, students also miss the opportunity to ask any follow-up questions to address information they may realize is missing when they go to write up their recommendation for the clients—something they would be able to do in a real-life scenario.
To address these limitations and support the professors’ pedagogical approach with emerging technology in the Law School classroom, Northwestern IT developed a generative AI chat simulation tool that allows the students to interview the two would-be co-founders—Andy and Dr. Gold—beyond the classroom.
Solution
The chat simulation tool mirrors a group text, a standard communication method between lawyers and their clients, and it allows each student to practice their interview skills after class.
When using the tool for the first time, students open a new chat with the AI clients and must (re)introduce themselves and follow up on the “previous meeting,” during which the professors acted as the clients in class. From there, the experience resembles a group chat: the student types a message to the AI clients, the AI clients respond, and they continue the conversation as needed for the assignment.
Northwestern IT worked with Law IT to compile information about Dr. Gold and Andy, including their backgrounds, roles in the new business, and future plans. They also gathered personality attributes for the clients. For example, Andy is very excitable and tends to be unrealistically optimistic about the business’s prospects, whereas Dr. Gold is more methodical and less comfortable taking risks. From this, the Northwestern IT team developed character profiles, enabling the AI to extrapolate responses to the students' questions. This ensured that the professors wouldn’t need to anticipate every possible question a student might have.
“We are always looking for new opportunities for our students to gain real-world skills that will help them hit the ground running as transactional lawyers. Text messaging is becoming increasingly important in practice, and we were excited that our students could practice this skill as they do work for their simulated clients, Andy and Dr. Gold,” said Esther Barron, clinical professor of law and director of the Donald Pritzker Entrepreneurship Law Center, Northwestern Pritzker School of Law.
Since human clients don’t always agree and can often change their mind, AI hallucinations and inaccuracies—traditionally considered limitations of AI systems—can add to the real-life experience of being a lawyer. One day, Andy might say he hopes to hire 50 people in the first month, but then he doubles that number to 100 the next day—this is part of the client management experience students will face as they become lawyers and need to navigate how to best glean information from their clients.
Screenshot image of student conversation with the chat simulation tool.
Impact
Once the tool was ready, it was deployed in a class of 40 students. Their assignment was to use the tool to interview their clients and then determine the appropriate financial entity for this new company. The tool does not grade the students; that job is still the professors’, so students include their conversations with Dr. Gold and Andy as part of their assignment submission, and Professors Reed and Barron review and provide feedback on their interviewing skills as part of the grading process.
The tool has been successful in recreating the characters' personalities and students have appreciated the freedom to engage with the characters after class ends.
“The Northwestern team that we worked with was amazing. They guided us through the process, understood our goals, and provided a great product. They seemed as excited as we were to create a meaningful tool for our students and to make sure the student experience would be optimized for learning. We are looking forward to deploying this in our other simulation classes as well,” said Stephen Reed, clinical professor of law and assistant director of the Donald Pritzker Entrepreneurship Law Center, Northwestern Pritzker School of Law.
Tech Specs and Future Use Cases
The tool consists of a web application written in React and an API written in Python, using the Flask framework. The AI characters are currently powered by GPT-4o, though the framework is flexible and the specific AI model can easily be swapped out, making way for future exploration and development opportunities. While the AI component of the project is central to its success, the user interface is also a key component for making the experience mimic a realistic exchange with clients rather than an interaction with ChatGPT. The tool’s integration with Canvas facilitates easy use in a classroom setting.
The Law School plans to continue the use of the tool, specifically in its San Francisco Immersion Program, which prepares students to work in the legal departments of Bay-Area companies. Soon, Andy and Dr. Gold won’t be the only AI clients that law students can interview. Northwestern IT will support Law IT in broadening the program for its faculty and students.
“Law IT is working with our faculty to bring additional AI application projects for the classroom and general use AI Assistants for the students’ law school careers. In addition, we are developing frameworks to scale the growth of support for AI use at Pritzker School of Law,” said Stephan Martone, manager of academic innovation, Law IT.
Northwestern IT is exploring other applications for the technology in other Northwestern schools and programs. This could be used by other schools where students need to develop strong client interviewing skills like business, psychology, journalism, medicine, and more.