In AI Battle, Trump Faces the European Front
As the White House tries to take the guardrails off of the technology business, Brussels could take the lead on policy decisions.
staff
JD Vance is certainly getting things done. In the middle of February, the vice president went to Munich to tell Europeans to stop isolating far-right parties, just after speaking at the Paris AI Action Summit, where he warned against strict government regulation. Talk about not knowing an audience: It would be hard to offend more Europeans in less time without kvetching about their vacation time.
This week, my colleague Kristin Robinson wrote a very smart column about what Vance’s — and presumably the Trump administration’s — reluctance to regulate AI might mean for copyright law in the U.S. Both copyright and AI are global issues, of course, so it’s worth noting that efforts by Silicon Valley to keep the Internet unregulated — not only in terms of copyright, but also in terms of privacy and competition law — often run aground in Europe. Vance, like Elon Musk, may simply resent that U.S. technology companies have to follow European laws when they do business there. If he wants to change that dynamic, though, he needs to start by assuring Europeans that the U.S. can regulate its own businesses — not tell them outright that it doesn’t want to do so.
Silicon Valley sees technology as an irresistible force but lawmakers in Brussels, who see privacy and authors’ rights as fundamental to society, have proven to be an immovable object. (Like Nate Dogg and Warren G, they have to regulate.) When they collide, as they have every few years for the past quarter-century, they release massive amounts of energy, in the form of absurd overstatements, and then each give a little ground. (Remember all the claims about how the European data-protection regulation would complicate the Web, or how the 2019 copyright directive would “break the internet?” Turns out it works fine.) In the end, these EU laws often become default global regulations, because it’s easier to run platforms the same way everywhere. And while all of them are pretty complicated, they tend to work reasonably well.
Like many politicians, Vance seems to see the development of AI as a race that one side can somehow win, to its sole benefit. Maybe. On a consumer level, though, online technology tends to emerge gradually and spread globally, and the winners are often companies that use their other products to become default standards. (The losers are often companies that employ more people and pay more taxes, which in politics isn’t so great.) Let’s face it: The best search engine is often the one on your phone; the best map system is whatever’s best integrated into the device you’re using. To the extent that policymakers see this as a race, does winning mean simply developing the best AI, even if it ends up turning into AM or Skynet? Or does winning mean developing AI technology that can create jobs as well as destroy them?
Much of this debate goes far beyond the scope of copyright — let alone the music business — and it’s humbling to consider the prospect of creating rules for something that’s smarter than humans. That’s an important distinction. While developing AI technology before other countries may be a national security issue that justifies a moon-shot urgency, that has nothing to do with allowing software to ingest Blue Öyster Cult songs without a license. Software algorithms are already creating works of art, and they will inevitably continue to do so. But let’s not relax copyright law out of a fear of needing to stay ahead of the Chinese.
Vance didn’t specifically mention copyright — the closest he got to the subject of content was saying “we feel strongly that AI must remain free from ideological bias.” But he did criticize European privacy regulations, which he said require “paying endless legal compliance costs or otherwise risking massive fines.” If there’s another way to protect individual privacy online, though, he didn’t mention it. For that matter, it’s hard to imagine a way to ensure AI remains free from bias without some kind of regulatory regime. Can Congress write and pass a fair and reasonable law to do that? Or will this depend on the same Europeans that Vance just made fun of?
That brings us back to copyright. In the Anglo-American world, including the U.S., copyright is essentially a commercial right, akin to a property right protected by statute. That right, like most, has some exceptions, most relevant fair use. The equivalent under the French civil law tradition is authors’ rights — droit d’auteur — which is more of a fundamental right. (I’m vastly oversimplifying this.) So what seems in the U.S. to be a debate about property rights is in most of the EU more of an issue of human rights. Governments have no choice but to protect them.
There’s going to be a similar debate about privacy. AI algorithms may soon be able to identify and find or deduce information about individuals that they would not choose to share. In some cases, such as security, this might be a good thing. In most, however, it has the potential to be awful: It’s one thing to use AI and databases to identify criminals, quite another to find people who might practice a certain religion or want to buy jeans. The U.S. may not have a problem with that, if people are out in public, but European countries will. As with Napster so many years ago, the relatively small music business could offer an advance look at what will become very important issues.
Inevitably, with the Trump administration, everything comes down to winning — more specifically getting the better end of the deal. At some point, AI will become just another commercial issue, and U.S. companies will only have access to foreign markets if they comply with the laws there. Vance wants to loosen them, which is fair enough. But this won’t help the U.S. — just one particular business in it. And Europeans will push back — as they should.
