NYT vs. LLMs
getting the decision right
Strictly enforcing the letter of what rights holders say in all cases would be crazy. Nullifying all rights and letting everyone do literal anything would also be crazy.
He is talking about copyright issues. He refers to the NYT lawsuit against Open AI, but the question of what is fair use comes up in many other instances as well.
About twenty years ago, I wrote an essay called “Legamorons,” meaning legal oxymorons. My claim was that strict enforcement of some laws would lead to chaos.
For example, suppose that the police suddenly decided to strictly enforce the speeding laws. Or suppose that laws against marijuana possession were enforced all the time under all circumstances. Strict enforcement of every law could lead to the vast majority of people ending up in jail or paying fines.
This is an uncomfortable situation. We want to be able to say that “the law is the law.” But in fact we cannot do that. As enforced, speeding laws do not apply unless (a) the police notice you and either (b) you are really being reckless or (b’) they have decided to collect a lot of fines in that place at that time.
In practice, law enforcement has a large discretionary component. The law is invoked only in court, where the attempt is to create a strict basis for settling the case.
There might be analogy with how some psychologists view human decisions and reason. They say that we decide what to do based on instinct, and then we use reason to justify our behavior. It might be that legal decisions are based on nuance, but then we use strict laws to justify the ruling.
I think that in the legal realm, we typically have two goals, which we are trading off against one other. We do not want to punish the innocent. And we do not want to let the guilty get away with it. The standard “beyond a reasonable doubt” means that we are more concerned about not punishing the innocent. But “beyond a reasonable doubt” does not mean “beyond all possible doubt.” Implicitly, there is still a small possibility that you are convicting an innocent person.
In the case of copyright, the two goals are: we want knowledge to spread; and we want people to produce knowledge. My way of putting this is: information wants to be free, but people need to get paid.
If large language models are directly undermining the business model of the NYT by making it possible for people to access articles without subscribing, then the providers of the models should have to obtain a license. But I don’t see the NYT’s business model as directly undermined in that way. So I would side with the defendants in the suit.
Substacks referenced above:
@
NYT vs. LLMs
My experience with enforcement of the speeding laws based on many years of driving long distances is that cops will typically pull you over if you exceed the speed limit by more than 10 mph {in some places a little more, in others a little less) whether you are driving recklessly or not. If the speed limits were enforced as posted (which today would not be technically difficult) drivers would quickly adjust, but there would be enormous public pressure to raise the limits to realistic levels.
Regarding the necessarily large discretionary aspect of law enforcement, this has been identified by the Left as means to advance their agenda. Billionaires lavishly fund the campaigns of radicals seeking office as prosecutors. The result of such political abuse has been not only great injustice, but a loss of public confidence in the law as fair and neutral. We have gone from trying to avoid wrongfully punishing the innocent to a Beria model; show me the man, I will find you a crime. We now avoid rightfully punishing the guilty as well, where they are favored by the Left. It seems likely that these evil practices will spread to the Right as a matter of self-defense.