:Harvey:’s Principles for AI Adoption and Rollout in Law Schools

Earlier this year, Harvey rolled out our Law School Program, through which we partner with top law schools to help students advance their AI proficiency. The program also aims to help faculty and staff incorporate AI into their curricula and consider more broadly the ways in which their institutions can prepare students, faculty, and alumni for the future of law.

Over the past several months we have worked with deans, librarians, faculty, and students across more than 20 schools. We hear distinct needs from each constituency: faculty want pedagogy that keeps reasoning intact, librarians want clear and secure governance, and students want permission to experiment with AI in the open. Harvey’s been able to play a unique role, understanding these unique perspectives and building them into our suggested roadmap for responsible AI implementation.

From those early conversations, it’s clear that people want to understand Harvey’s point of view and how schools should leverage AI, as well as where schools should actively set limits and boundaries on AI usage.

We view this as part of our broader commitment to responsible AI usage, and critical to earning and building trust with organizations for whom academic integrity is foundational to everything else they do. To that end, below we’ve outlined Harvey’s principles for AI adoption alongside suggested use cases where many schools have seen promise starting out.

These principles articulate Harvey’s framework for helping law schools integrate AI thoughtfully into legal education while preserving the rigor, ethics, and human judgment at the heart of the profession.

Reasoning has a sacred place in legal pedagogy, and AI should support — not substitute — the cognitive steps of legal reasoning. Just like a good tutor engages students rather than giving them the answer, AI platforms should allow students to cut through noise with a thought partner.

We believe in human-in-the-loop AI usage, meaning the user plays a critical role in verifying AI output and confirming the accuracy, jurisdiction, and relevance before submitting a brief or assignment.

Skills like prompting will become core to completing legal tasks with AI, so giving students the opportunity to learn how to apply structured inquiry to prompting and design workflows with AI will lay the foundation for expertise that will help them in their legal careers long term.

The honor code remains the honor code, and your institution’s AI policy should be clear to faculty and students, even within software platforms, so everyone has clear insight into the responsibility they bear for honoring those standards and commitments.

Providing equal and free access to AI tools as part of students’ education helps level the playing field for everyone to learn critical skills associated with AI fluency and increase comfort using it in a legal setting. Institution-provided access under school governance also yields greater security and accuracy. When students have a secure workspace and clear norms, they learn responsible habits rather than turning to public tools that lack privacy or accuracy.

Principles are helpful in understanding our approach, but many institutions want concrete examples of how other schools are putting this into practice. Below, we’ve summarized our learnings to date into four recommended stages for rollout.

Each law school should publish a clear, public stance on responsible AI use that aligns with its honor code and academic integrity rules. If appropriate, have students acknowledge a brief responsible-use addendum before using generative AI tools.

The goal is not to rewrite policy, but to clarify expectations for this new context. When students and faculty understand the boundaries, AI becomes a tool for open learning, responsible experimentation, and trust across the institution. Onboarding sessions for new AI tools should also include clear articulation of hallucination risk, confidentiality, and verification.

Adoption accelerates when faculty see AI tools as a teaching aid to help students both learn the substance of their course material and get familiar with how lawyers are using AI in practice. We often see the highest value return from faculty members delivering clear, structured assignments to help students get comfortable with proper real-world use cases for AI.

A good example here might be assigning an initial first draft that is fully manual, then making the second component of the assignment an AI critique and student verification of that critique. A final component could have the student reflect on what the AI tooling helped them improve, and where the AI may have missed. Grading in this instance could be predicated on reasoning advancement, prompt utilization and refinement, and accuracy.

Where we’ve seen success in this space is schools giving students explicit permission to use AI tools to augment study habits. Some examples of use cases that have been helpful to students to date include:

Every geography has its own nuances with legal education, and every institution has its own approach to pedagogy and teaching each element of the law to students. Given that, it’s imperative that students and faculty have opportunities to share feedback: with each other, with the school around their learning experience, and also with Harvey regarding product experience and user onboarding. Leveraging AI in the legal education space is a new skill for everyone, so it’s imperative to create feedback loops that allow us to identify opportunities and challenges early on and iterate on them together.

Law schools have a profound impact on shaping the future of the legal field. My own experience with law school faculty reflected the highest commitment to ethics and legal reasoning, while at the same time ensuring students are prepared to create the future versus the past. Shaping the future means upholding the highest standards of legal integrity while also embracing new tools and approaches that enhance efficiency, effectiveness, and impact.

This places a significant amount of pressure on schools to balance rapid innovation with technology with consideration and clarity on responsible AI to remain competitive. At Harvey, our goal is to be a thought partner in this endeavor and our hope is that these principles and frameworks will help make it easier for schools to choose and adopt AI in legal education moving forward.