How Should the U.S. Regulate AI?

How should the United States regulate artificial intelligence (AI)? And does regulating American AI impact competition with China? Timothy Wu, Julius Silver Professor of Law, Science and Technology at Columbia University, argues, "It's not crazy to have terms. So if you're going to fund a university or a private institution doing the research, you know, ultimately you don't want government to be giving them huge amount of money so they can make more money. It's a form of corruption. So you have to avoid that. They have to share it with the public, unless it's sensitive. And then you might want to say...be very careful when you're doing research that could be unethical or highly dangerous. You know, create the robot uprising....For one thing, if what you're requiring is that the research is shared, you know, it's made it available to other scientists, you're actually promulgating a faster approach. In that case, not regulating would be slowing down AI, because you're letting one company keep it all secret. So there's one factor. Second...if you're concerned about the future of civilization or humanity from creating dangerous AI, so you urge some caution on that, and if that gives you a disadvantage, well, I'll take it."

This interview from the FPA archives was recorded in 2019.