You’ve probably felt it—the quiet anxiety that the rules governing our professional lives are written by people who don’t understand the tools we actually use.
Look no further than UC Berkeley School of Law’s new Artificial Intelligence Policy. The school has attempted to draw a bright, definitive line between human and machine intelligence. It is a massive, predictable failure.
The policy states that AI is strictly prohibited for any purpose in any exam situation. Read literally, this bans hearing aids. Modern hearing aids use neural networks and transformers for noise reduction. The rule is so desperately broad that it accidentally disables basic accessibility, yet so intellectually narrow that it completely misses the existential threat to the legal profession.
You can’t legislate away a technological revolution; you can only guarantee your students are unarmed when it arrives.
We are told these blanket bans protect academic rigor. They don’t. As one observer perfectly noted in response to the policy: ‘You have to be jacked to use the jackhammer.’ Berkeley isn’t protecting rigor. They are avoiding the responsibility of building students into the kind of strong, discerning operators who can actually wield AI with judgment.
Instead of teaching students how to supervise a machine that hallucinates case law, the faculty has chosen to plug their ears and pretend the machine doesn’t exist. They permit AI for ‘source research’ but ban it for argumentation, pretending a firewall exists where there is absolutely none.
Pretending a linguistic engine won’t disrupt a linguistic profession isn’t policy—it’s professional denial.
Law is built on established arguments, gray areas, and linguistic subtleties. It is, at its core, advanced language manipulation. If that is the crux of the profession, law is far more automatable than any dean is willing to admit. The elephant in the room is that the junior associate’s job is already on life support.
The fortress walls have already been breached. The AI is inside the walls, in our phones, in our hearing aids, in our research databases. Writing policies that pretend otherwise doesn’t protect students. It leaves them swinging a hammer they don’t know how to use, while the world moves on to jackhammers.
FAQ
Q: Doesn't banning AI ensure students actually learn the foundational skills?
A: No, it ensures they learn skills for a world that no longer exists. You don't teach foundational judgment by hiding the tool; you teach it by forcing students to use the tool under strict supervision.
Q: What should law schools do instead of banning AI?
A: Treat AI like a power tool. Make students use it, break it, and verify its outputs. If a student can't instantly spot a hallucinated case citation, they shouldn't pass. That’s the new standard of rigor.
Q: Is law really just language manipulation that AI can easily replace?
A: A massive chunk of it is. The policy admits AI is good enough for research, but pretends legal argumentation is sacred human ground. It's not. If an AI can write a better brief, the barrier isn't intelligence—it's just regulatory capture.