Every law is unenforceable

Against the learned helplessness around technology

This morning, I stumbled across a post bemoaning UC Berkeley Law School’s near-total ban on AI use for school purposes. The reason? It is “unrealistic” and “unenforceable.” Instead of fighting the tide of technological progress, Berkeley should figure out a way to fit their pedagogy with the surrounding reality.

We see this again and again. Regulating technology is “unrealistic” because it would be “unenforceable.” And yes, it is the case that online technologies make it exceptionally difficult to track and ban online-mediated activities. Outright banning Uber driving or Airbnb rentals would have been difficult to achieve in practice. Outlawing crypto mining would present similar difficulties. As for the use of AI in school, I personally ban all use of AI—even for “brainstorming,” because dear Lord, surely coming up with ideas is a core human competency!—and I’m sure that I have missed papers that used AI for various purposes and even included AI-generated text.

What the hot takes about “unenforceability” fail to note is that that is true of literally every law, ever. Murders and rapes and robberies take place every single day, in every country on earth. Drunk driving is rendered virtually inevitable by the car-centric design of most American communities. Short of instituting an impossible dystopia of omnisurveillance, there is little that law enforcement can do to directly prevent a drunk person from getting into the car or a person in their private home from picking up a knife to stab their partner or a self-selecting person from accosting a random stranger on the street. Indeed, even in the dystopia, people would likely find a way! And yet no one argues that the behaviors I mention in this paragraph should be legal. No one argues, for example, that maybe if you’re concerned about drunk drivers, you should drive more defensively instead of depriving someone else of a way to get home from the bar. No one says that drunk driving is inevitable and here to stay and we just have to live with it.

That is because, crucially, drunk driving does not have a website. If it did, then presumably everyone would be standing ready to chime in that a ban on drunk driving is “unenforceable”—in the sense that no ban could be 100% effective in eliminating the behavior—and hence it should not even be attempted. That’s because websites exist beyond the law. If you have a website, you have a direct connection to the very pulse of historical necessity. In a sense, you are your own law. If you have a website, you can completely upend the transportation system. If you have a website, you can convert whole neighborhoods of a major city into an illegal hotel chain—an action that represents simultaneously the deepest expression of freedom (people are choosing to put those units on the market you have created!) and destiny (the good old ever-changing world of technology marches on!). If you have a website, you can radically change people’s patterns of social interaction and create epidemics of political radicalization and mass shootings. If you have a website, you can basically render education impossible. You can even give young people step-by-step instructions in how to commit suicide and convince them to stick to it if they express doubts.

None of this can be stopped, because websites are all-powerful. What’s more, they deserve to be all-powerful, in large part because everyone wants them to be all-powerful. Be honest—you’ve used a website before, haven’t you? So you know how it is. Life without websites is unimaginable. If you try to shut one down, it can just move its server and you type a different thing and, bam: the website is back, baby! It’s good again! How could anyone fight such a thing? Why would you even want to attempt to interfere in such technological excellence?

So yes, people should be allowed to graduate from law school after letting a website do all the work for them. They should be competing for jobs with the advantages that degree bestows, without having done any of the work that the degree is meant to symbolize. Doing anything else would require standing up to the almighty website! Why would you do that? And why would you even try?

I can think of a reason: because cheating is wrong and people should feel bad if they do it. Even if we cannot prevent all use of AI, the rule requires students to hide their use. It prevents them from discussing it as though it is normal, because then they will face severe consequences. That combination of factors may even induce them to feel the kind of shame that is normally associated with activities that you have to hide in order to avoid punishment and social sanction. If someone wants to waste their opportunity for education by using AI, they should be forced to act like someone who should be ashamed of their actions, because they should in fact be ashamed of their actions. They should be aware of the possibility that they will have to spend the rest of their life admitting to people, “I was accepted to one of the most prestigious law schools in the country and I gave up all the opportunities that represents because I let the fucking website do my homework and someone caught me.”

In other words, sometimes you make a rule against something because you think that thing is wrong and you want to send a clear message that it’s wrong and people should be ashamed to do it. That’s a thing that people who live in a society have been doing from time immemorial. It is the right and duty of every human community and every human institution to make rules that express its values.

The faculty of UC Berkeley Law School have exercised that right and fulfilled that duty by sending a clear message that the use of LLMs as a shortcut is incompatible with their values. I fully support and congratulate them on this obviously correct stance. I’m willing to accept that there may be narrow technical contexts where the use of AI technology is legitimate, but the general meta-skill of assimilating and digesting information is not and can never be one of those contexts.

The website cannot understand for you, and it will never be able to. In recognition of this fact, faculty members everywhere have a duty not to let the website steal away their students’ one chance at education. We should make every effort to prevent that from happening instead of giving up in advance because we’ve imposed the impossible and absurd standard of 100% effectiveness on ourselves. And if you think otherwise, I urge you to unsubscribe, because I don’t want to be associated with you—though of course, since a website is involved, I could never enforce such a suggestion.