In the Phaedrus, Plato tells us that Socrates distrusted the written word. He believed that committing knowledge to writing would weaken memory and discourage dialogue. Writing, he warned, is inferior to debate because it cannot respond. It only creates an appearance of wisdom, not genuine wisdom.
The parallel with AI is clear. We now have systems that write, summarize, argue, program, and draw at a speed and with a breadth of knowledge that no human being can match. They make us feel competent and informed without us first having to face the discomfort inherent in any learning process. The danger is not that AI will think, but that we will stop thinking for ourselves.
However, the history of writing offers a less pessimistic alternative. As writing became established as a central element of education, oral debate was not abandoned. What was essential in debate was preserved, to the point that we still use the Socratic method in teaching today, especially in law schools, because we know that questions posed in the classroom teach students to think quickly and develop judgment in a way that no text can replace. At the same time, we learned new ways of thinking. Silent reading, once unusual, became second nature. We learned to reason internally and to engage in a dialogue with the written text that has allowed us to continue conversing with Socrates even today.
In other words, we don’t choose between debate and text. We broaden our ways of thinking and, in doing so, become better thinkers.
We now face a similar threshold with artificial intelligence. Just as literacy became a fundamental requirement for participation in society, AI literacy is becoming essential for informed engagement in law, public life, and democratic decision-making. A law graduate entering the legal profession without the ability to understand and critically work with AI, to question its assumptions, and to consciously address its ethical limitations will be as ill-prepared as a lawyer who cannot read. A legal education that excludes AI is, therefore, an incomplete education.
If legal education fails to adapt, we risk losing a generation of students trained with pre-AI methods for a post-AI world. Their ways of thinking, learning, and writing will have been shaped by these tools long before they reach our classrooms, and the professions they enter will demand their critical and responsible use of them. Ignoring AI will not make our students think better. It will lead to demotivation and frustration, while leaving them ill-prepared.
This isn’t just about our students’ careers, however important that may be. The consequences of our inaction don’t fall solely on our graduates. The law influences our daily lives, often silently but ubiquitously, and when technology transforms how society functions, the law must transform with it. It will be our students, as future lawyers, judges, policymakers, or law professors, who must translate these changes into fair rules. If we don’t prepare them for this, the gap will manifest itself in every other area.
So what should legal education do? It’s not enough to simply add a subject or adopt a single rule. Certainly, it’s pointless to turn a blind eye and prohibit the use of AI. The challenge is more profound. We need to strengthen the foundations while simultaneously evolving with new developments. It’s not about choosing between tradition and innovation, but about finding a balance between the two.
We must redouble our efforts to cultivate what makes us human: intellectual curiosity, purpose, creativity, the drive to question, and critical thinking. These dimensions become even more important when AI performs a substantial portion of the technical work.
When nearly everyone has access to systems capable of writing reports, summarizing cases, or outlining legal arguments, the key questions change. What will students choose to do with that capability? What questions will they ask? What issues will matter enough to them to go beyond a merely acceptable answer?
AI cannot decide these matters. It cannot, and should not, determine students’ motivations or priorities. The ability to decide what we want, to define our purpose, and to pursue our goals remains central to our capacity for action as human beings.
However, we must evolve. If we truly aspire to integrate AI as a resource that fosters the intellectual development of our students, rather than undermining it, at least five major changes must occur.
- We must educate ourselves.
Professors and academic leaders must understand AI both functionally and conceptually: what these tools are, how they work, what challenges they pose, and how they can be used both to teach law and to improve its practice. You cannot teach what you do not understand.
- All students must take AI literacy courses.
AI literacy is not optional, especially not in law schools. It must be universal, rigorous, and structural. Students must learn how law, technology, and society interact; how AI systems reason and why they fail; how they influence reasoning and judgment; and what their responsible use entails in practice, both in everyday life and in professional practice. Students should not rely on readily available answers they don’t know how to evaluate.
- We must create spaces for simulation, reflection, and metacognition.
Students need environments where they can openly work with AI to simulate current legal practice. They need to experiment with tools, but also analyze the results, identify errors, and reflect on their own thinking as part of the learning process. This requires access to resources, dedicated coursework, and structured practice.
- We need to decide in which areas AI should not be used.
We must consciously consider how to design tasks that avoid delegating cognitive processes to AI and the resulting erosion of human intelligence. It is necessary to preserve memory and attention, in-depth reading, reasoning, critical thinking, and ethical frameworks. Likewise, it is essential to strengthen the teaching of basic legal skills and ensure that students learn to perform them without AI assistance, such as detecting problems, applying precedents, identifying and mitigating risks, and defending a legal position.
- We must transform our teaching methodologies to raise the intellectual level we demand of our students.
This implies changing assignments, exams, and classroom dynamics. If AI is capable of performing specific tasks, these can no longer be the ultimate goal of learning. It is essential that students learn to navigate the interpretive, strategic, and ethical aspects of legal reasoning. These dimensions must be integrated into the methodology, not added afterward as a belated resource.
Change is our duty. We must do it because the legal profession is changing, and because we have an obligation to our students: to prepare them for the world they will soon have to face, not for the world in which we were trained.
The written word did not diminish our capabilities; it transformed them. Human beings adapt. That is our genius. And, once again, the time has come to put it into practice.
By Carmen Perez-Llorca Zamora, Vice Dean of IE Law School
Article originally published in English on IE Insights https://www.ie.edu/insights/articles/ai-legal-education-and-the-duty-to-change/
