OpenAI's newest reasoning model, o3, has achieved something remarkable: it scored three A+ grades on actual law school final exams when graded on the same curve as human students. Seven law professors at the University of Maryland administered their Spring 2025 final exams to the o3 model with its "reasoning effort" parameter set to high, revealing a dramatic leap in AI's ability to handle complex legal reasoning. Three semesters earlier, the older GPT-4-turbo model would have earned grades between B+ and D on the same exams. The o3 model's results, one A, one A-, two B+s, and a B across the remaining exams, underscore how rapidly AI reasoning capabilities are advancing.

Why Are Law Schools Suddenly Worried About AI Competence?

The o3 performance on law school exams arrives at a critical moment for legal education. Law schools are grappling with a fundamental tension: generative AI tools are becoming essential to legal practice, yet most students graduate without foundational skills to use them safely and effectively. A recent framework published in June 2026 highlights the urgency of this gap. According to research on the topic, gen AI literacy for new associates is now expected, not optional, yet most law schools have not caught up with curriculum changes. This creates a risky situation where young lawyers enter practice unprepared to leverage AI responsibly, or worse, to recognize when AI outputs are unreliable.

The stakes are high. Throughout 2025, courts imposed judicial sanctions on attorneys who submitted AI-hallucinated case citations, and judges grew increasingly frustrated with AI-generated filings containing fabricated legal references. Law schools are now shifting from restricting AI use to requiring AI literacy as a core competency. The question is no longer whether students should learn to use AI, but how to teach them to do so without eroding their foundational legal reasoning skills.

Does Using AI Early On Damage Legal Reasoning Skills?

One of the biggest concerns among legal educators is whether students who rely on AI early in their learning process will develop weaker independent reasoning abilities. A rigorous empirical study published in April 2026 tested this directly by having upper-level law students complete legal writing tasks with and without AI assistance. The results were more nuanced than either AI enthusiasts or critics expected. Students who used AI to help craft synthesis memos produced substantially stronger work and completed the task more quickly. Contrary to what researchers had initially hypothesized, AI exposure at this initial stage did not diminish downstream comprehension of the underlying legal principles. In other words, using AI early did not appear to hollow out students' understanding when they later worked without it.

However, the study also emphasized that the effects of AI depend critically on when and how students use it. The research suggests best practices for AI integration rather than blanket approval or rejection. This finding has profound implications for curriculum design: law schools can incorporate AI as a learning tool without sacrificing the deep analytical skills that form the foundation of legal practice.

How to Integrate AI Literacy Into Law School Curricula

• Coordinate Existing Resources: Law schools need not overhaul their entire curriculum or make major investments. Instead, they can coordinate what they already have by ensuring every student graduates with foundational gen AI literacy skills, using a framework built on three co-equal questions: what students learn, when and where they learn it, and who coordinates the learning.
• Teach Foundational Skills First: First-year legal writing courses should maintain the student as the actual author of work product, allowing them to develop essential judgment, analytical abilities, and writing skills that form the foundation of the legal profession. Only after learning how to do the work themselves can future practitioners properly and effectively utilize AI as a tool.
• Introduce AI With Thoughtful Guardrails: As students cultivate core skills, legal writing professors can introduce gen AI as a tool to potentially enhance efficiency and effectiveness in certain circumstances, but with clear guardrails in place to prevent over-reliance or misuse.
• Develop Clear Institutional Policies: Law schools must establish clear governing policies for AI use and adoption in legal education, moving beyond ad-hoc approaches to systematic frameworks that address the evolving role of AI in legal scholarship and practice.

What Do Law Schools Actually Need to Teach About AI?

The challenge extends beyond just teaching students how to use AI tools. Law schools must also educate students about the risks and limitations of generative AI in legal work. Throughout 2025, the legal profession witnessed a wave of judicial sanctions imposed on attorneys who submitted AI-hallucinated case citations. These incidents revealed a critical gap: many lawyers did not understand that AI language models can confidently generate false citations that sound plausible but do not exist. Law schools now recognize that teaching students to verify AI outputs, understand hallucination risks, and maintain ethical compliance is as important as teaching them to use the tools in the first place.

The broader question facing legal academia is how to maintain scholarly integrity while embracing AI as an integral tool. Law reviews, which shape legal scholarship, have largely lacked clear AI policies. Emerging guidance suggests that rather than drawing rigid lines between human and machine contributions, legal academia should promote a better understanding of how AI can complement human authorship. The key to maintaining scholarly integrity is not restricting AI's use but promoting transparency and thoughtful integration.

The o3 model's success on law school exams is a watershed moment. It demonstrates that AI reasoning capabilities have reached a level where they can handle complex legal analysis at a high standard. For law schools, this is both an opportunity and an urgent call to action. The institutions that successfully integrate AI literacy into their curricula, while maintaining rigorous standards for independent legal reasoning, will graduate lawyers best prepared for the future of the profession. Those that delay risk sending unprepared graduates into a legal landscape where AI competence is no longer optional.