Roughly three years after ChatGPT’s release, law schools across the world have started to draw lines lines on AI use. One of the first was UC Berkley, followed by University of Texas at Austin sometime after.

This July, the University of Chicago Law School announced it will ban laptops, tablets, and phones in the classroom. Last week, the University of the Philippines (UP) College of Law announced it will be implementing a similar restrictive policy for freshmen classes starting this semester (AY 2026-2027). 

The policies advanced by these law schools seem aligned with the Supreme Court’s own recent reminder on the use of AI by lawyers and judges. Their common focus is re-emphasizing the human element of justice and the responsible use of tech. UP Law’s policy prohibits “the use of digital or electronic devices (e.g., phones, laptops, tablets, smart watches, smart glasses, recorders)” for first year classes (which is regulation of, not a total ban on AI). 

I didn’t have a hand in formulating UP Law’s policy. This is not to say I don’t agree with it. For decades I’ve never allowed electronic devices inside the classroom except for very specific (and rare) reasons. What follows is part of the thinking that underpins this long-standing rule of mine. It shouldn’t be attributed to UP Law whose faculty I don’t have the authority to represent. 

Why do we still teach children how to add and subtract when calculators already exist? Because learning is not just about obtaining the correct answer. How we get it matters. In cognitive psychology, we conduct experiments that show how repeated practice of a task strengthens the ability to perform that task. These fall under neuroplasticity and skill formation. Which are fancy ways of saying there is no substitute for practice and repetition.

First year law students need to learn a different language (legal gobbledygook), a different way of writing (legal jargon), and a different way of thinking (legalese). This requires the brain to form new neural pathways. A ponderous process, but well worth it. By their 3^rd or 4^th year, law students notice how fast they can absorb cases. What used to take six hours (ie. Javellana), they can do in less than one-fourth of the time.

If a student outsources tasks to AI, then it’s AI that routinely performs the (boring) intermediate steps needed for skill formation or memory. This is a form of cognitive offloading. When areas of the brain are not actively used, they tend to deteriorate over time. 

Thinking on their feet 

Future lawyers need to develop the ability to think on their feet — to answer without external aid and in a short span of time. Consider the ongoing impeachment trial of the Vice President. Objections are raised and answered within seconds, not minutes. Polished Q and As (trial scripts) collapse within the first barrage of objections. 

Corporate practice presents similar challenges.

Our lawyers in the firm encounter high-pressure meetings, critical negotiations and bet-the-company conversations where there’s no space to check one’s laptop or phone because the general counsel or CEO is in front of you waiting. Of course, not all legal tasks are like this but, we train for the exigencies, not race to the bottom.   

There’s also the ethical reason. In academic settings, when a student submits something to the professor or teacher, it must be the student’s product, not someone else’s. Law practice takes this to a higher level. When you sign something, you tell the court, the partner or the client, “this is my work, I stand by it.” This matters because the legal profession ultimately relies on trust.

A client goes to a lawyer and shares his problem with the expectation that the lawyer will use his skill and analysis to provide the answer. The sanctity of this attorney-client relationship is the source of most of a lawyer’s privileges. When a lawyer outsources analysis to a machine, the fundamental premise of the attorney-client relationship is altered. Trust is shaken. 

There are cases where AI can be of great help to legal education and practice. This happens when AI augments rather than substitutes. A lawyer with years of practice gets a markedly different experience than a first year law student when using AI. A veteran detects errors, nuance, or hallucinations. AI thus augments his existing knowledge and skill (when it gets things right), producing efficiency. When a person lacks the foundational legal skills and knowledge to scrutinize what a chatbot regurgitates, the answer effectively substitutes for it. 

‘Perfect’ answers

To be honest, what really worries me about the prevalence of AI misuse is its uncanny ability to provide “perfect” answers.

AI has gotten so good at getting the answer right, students are inadvertently skipping that part of the learning process that provides the greatest potential for growth. Failing. Success doesn’t teach wisdom. A student who gets it right all the time leaves little room for self-check and reflection. He never finds the need to sort out where he got things wrong. 

Are these law schools just being “Luddites”? To say that would be ironic. The history of this term isn’t about ignorant fear of progress. It’s about abusing technology, oppression, and displacement of workers for more profit. Others say it robs freshmen of the chance to “shape the future [of AI].” Perhaps shaping the future rests on hard work, experience and, solid career choices, not whether you were allowed unfettered use of Gemini during your freshman year in law school. 

I don’t claim to understand AI deep enough to know how it will all turn out. I am just a teacher. Unfortunately, a good number of students aren’t using it “on a higher level.” They are using it to write essays, create case digests, or to game recitation. Will one of my students turn out to “shape AI”? That’s possible. Which is why they’re free to interact with it however they want (as they have done so for years before entering law school). I only ask that they do it outside our class hours. I doubt if their destiny will be defeated by unplugging them from their phones (or laptops) a few hours per day. 

Is this the right bet? Only time will tell. But if we are to be frank, it is no less wise than those who want the reverse (ie use AI as much as you want). A decade ago, we flooded classrooms with tablets and laptops. Now, we are taking them back. Law firms are wondering if the acclaimed efficiency gains are worth it since now, senior lawyers spend time double-checking individual sources for accuracy or worse, hallucinations. Structured AI use can be valuable. That can come after we teach freshmen the basics.

Three years ago, within days of its public release, I introduced ChatGPT to my students. I eagerly shared how it will cut study times and make things easier. One of them couldn’t resist asking, “But sir aren’t you afraid we’ll use it?” I smiled and said, “Some of you already are. So, this is my way of saying that there are risks. Don’t get addicted to easy things.” I still mean that.

The lawyers who will shape the future of tech are those who are not  beholden to it. – Rappler.com

John Molo teaches in UP Law and UA&P School of Law and has argued several landmark cases before the Supreme Court.