The Artist and the Algorithm
A conversation between Joseph Gordon-Levitt and Andrew Nusca of Fortune
Interview Transcript:
Andrew Nusca: Thank you for being here. All right, good morning, everybody. I know, right? Can you feel the caffeine flowing? So you were giving an interview last month and you said, “I love new technology.” Are you a tech optimist, a tech pessimist, a tech skeptic?
Joseph Gordon-Levitt: Well, I do hold quite a lot of optimism for AI. I think it could be so great in so many ways. But I also think there are probably potential pitfalls, and we’re running right towards them in many ways. And I think that if we let this technology unfold driven purely by business incentives, and there are no guardrails put there for the public good, we could see some really dark outcomes. It’s what we’ve seen with a lot of tech over the last 20 years. We’ve seen it with social media, which was so good in so many ways. It’d be hard for you to find someone who is more positive and optimistic about social media than me. I did this online community for creativity using social media tech called HitRecord. But obviously, we’ve also seen all these damaging side effects, and maybe we could have avoided some of those with some proper guardrails. Let’s hope we don’t make those same mistakes again.
Andrew Nusca: Yeah, indeed. So you have been quite out there about these guardrails, this need for policy change. I mean, there’s a whole litany of things that we can do to solve for this, and hopefully we’ll talk about it today. You signed a petition for safer superintelligence, as we’re calling it now. You’ve been rather, well, supportive of one and more critical of another, AI bills here in the state of California. And you have definitely been critical of the chatbots by Meta and other companies that have verged into, shall we say, erotic content. Yes?
Joseph Gordon-Levitt: For kids.
Andrew Nusca: For kids. Erotic content for eight-year-olds. Are you in favor of erotic content for eight-year-olds?
Joseph Gordon-Levitt: Yeah, exactly. Because that is what the chief ethicist of Meta approved in their AI companions on their platforms.
Andrew Nusca: So, you know, we’re in a moment right now where it feels like all of the wins, from the White House, from the states, certainly from the tech companies, everybody’s just trying to... Well, it’s framed in a sort of global competitive way, right? No one’s saying, “Well, we’re going to make more money.” It’s more just like, “We have to get there first.” But, of course, all of the concerns that you have are being kind of pushed to the side in that. So what do we need to do? Should we talk head-on about that key argument that they make?
Joseph Gordon-Levitt: Oh, “we have to get there first.” I mean, it turns very quickly into this geopolitical thing.
Andrew Nusca: Correct. We have to beat China.
Joseph Gordon-Levitt: That’s right.
Andrew Nusca: Okay. So what do we do about that? Because clearly, that’s where the winds are blowing.
Joseph Gordon-Levitt: So this is a very powerful stroke of what I would call storytelling. This is what I’ve spent my life working on, storytelling. And you’ve got to unpack the layers of the story when they say, “We’ve got to beat China.” The implicit, and actually sometimes explicit, comparison there is to the race to nuclear weapons during World War II. There was this moment in history where the U.S. mounted the Manhattan Project, and the U.S. was racing against the Nazis to take Einstein’s theory and build a bomb with it. Whoever would get there first and push that button was going to win the war and kind of decide whether the world was going to become fascist or liberal. Not liberal in today’s sense. And that story is told today as if the same thing is going to happen. As if there’s this thing. They use terms like “superintelligence.” You just said they also use a term called “AGI.” There are no definitions for these terms. You can ask the most informed engineer or policy person in AI, and I have, and they’re like, “Yeah, there’s not really a definition for this.” Because it’s all a story. It’s hand-waving. It’s a sleight of hand. There is not going to be a moment where they push a button and all of a sudden win the war and decide that America is going to dominate instead of China. That’s actually not coming. But if you try to say, “Hey, maybe we need to put some guardrails in place. Maybe instead of just accelerating towards your maximum profits, we need to steer. We can still go fast, but let’s build stuff that’s going to also be good for everybody, not just for profits,” the first thing they say is, “Yeah, but then we’ll lose to China.” But it’s a sort of paper-thin story that doesn’t really make sense.
Andrew Nusca: This is an unfair question, but do you feel that artificial intelligence as it is framed today is more dangerous or less dangerous than nuclear weapons in the middle of the century?
Joseph Gordon-Levitt: Oh. Obviously, the dangers are very different. But yeah, look, I’ve spent the last, because you mentioned I’m about to direct a movie about this stuff, I’ve spent the last several years talking to all kinds of experts and engineers and policy people and people who work in the big tech companies and people who work in academia and everyone in between, governments. And there’s no agreement on this. So I don’t know. I don’t know the answer. There definitely are a lot of people, a lot of experts. There’s a survey that I think half of published AI researchers give it a one in 10 chance that something extremely catastrophic happens with AI, along the lines of, you know, nuclear warfare or whatever. I don’t know. I certainly don’t know. But if these experts are giving it a one in 10 chance, and it’s not just a few of these experts, it’s like half of the experts, that sounds pretty bad to me. They make the analogy: if half the avionics engineers said, “I’d give that plane a one in 10 chance of crashing,” would you get on the plane? No, you probably wouldn’t get on the plane. So I don’t know. I mean, I think those concerns are real. I think there are also a lot of other related but important concerns that might not be so catastrophic as that, but get really dark and dystopian if you take them down the road a few steps into the future. And those are important to talk about too.
Andrew Nusca: Sure, sure, sure. I’ve seen enough sci-fi movies and television shows to know that it can get pretty dark and dystopian. It’s a funny thing, right? And speaking of movies, this is one of the things that the folks who don’t want any laws will say: your concerns are science fiction.
Joseph Gordon-Levitt: But the real people now, the real ones making this stuff, are saying, “This isn’t science fiction. This is really happening.” So again, there are all these evasive tactics that I think are thinly veiled attempts to just protect their bottom line.
Andrew Nusca: Sure, sure. And I will just say, I don’t buy that argument either, because let’s face it, most of what we saw in Minority Report all those years ago is kind of real now. It’s remarkable to me that with every passing year, we manage to knock off one more of those technologies that seemed so futuristic at the time. All right, so. You know, there have been attempts. One of the other things that you’ve come out against is the notion, because right now it’s just a notion, of a federal suppression of states’ ability to regulate AI. The White House has made its position clear on that. How it goes about doing that, that’s a matter of debate. It’s tried to slip in a couple of passages in, what, two bills now? It tried in the big, beautiful bill.
Joseph Gordon-Levitt: Correct.
Andrew Nusca: And then the other one. Then it just tried in the, what do they call it, the National Defense Authorization Act or whatever, like the yearly procedural budget passage for the military. They tried to sneak it in there.
Joseph Gordon-Levitt: Exactly.
Andrew Nusca: And so that’s not happened. But the latest is that Trump wants to just, as he is wont to do, put through an executive order kind of doing that now. Setting aside the legal support of that, you’re not a lawyer and neither am I, what’s your take on this? Because obviously you’re against kind of clearing the path for AI. You want some of those speed bumps. But here we are, finding a Republican White House in tension with Republican states even.
Joseph Gordon-Levitt: Yeah. I was just in Utah.
Andrew Nusca: Yes. We have a wonderful Brainstorm tech conference there.
Joseph Gordon-Levitt: Right on. I was sitting at a dinner with a whole bunch of Republicans and just agreeing left and right about how important it is to make laws that govern this technology. For many of them, it’s around kids, right? It’s supporting kids.
Andrew Nusca: That’s definitely an alignment on the right side.
Joseph Gordon-Levitt: It’s about kids, but it’s also for Republicans who are in favor of things like competition. Sometimes you need laws to make sure that there is a fair playing ground for business and competition to flourish. So there was a lot of talk about that too. And Utah has actually done more than just about any other state to regulate AI. And besides right-wing and left-wing lawmakers, I think the public, left, right, and center, and I’ve looked at polls about this, they want regulation on AI. They don’t trust these big businesses to do what’s in the best interest of normal people. They know that these businesses can’t do that. And it’s not because the businesses are full of bad people. Not at all. It’s because if you’re a business, and everyone here at the Fortune event probably knows this better than I do, if you’re a business, you have to try to do good business. If you try to prioritize the public good, if you try to take the high road, you’re just going to get beat by a competitor who’s taking the low road. You can’t really prioritize the public good if you’re a business. That’s why there needs to be an interplay between the private sector and laws. Every other sector that impacts people’s lives has laws. Whether it’s the companies building cars, or the companies who have banks, or the companies making medicine, whatever, everybody has to follow some laws. Why should the companies building this technology not have to follow any laws? It doesn’t really make any sense.
Andrew Nusca: Indeed. We even have international laws about the use of nuclear force. So, you know, cheers.
Joseph Gordon-Levitt: Indeed.
Andrew Nusca: You posted on your social media recently something that resonated with me quite a lot as the father of two boys. Now, my boys are young. I have a five-year-old and a three-month-old.
Joseph Gordon-Levitt: Thank you.
Andrew Nusca: As quite a few of you know, because he was born in the middle of the last Brainstorm that we had. It was about artificial intelligence in toys. You have three kids. I can already feel the dad energy coming off of you on that subject. But tell me a little bit more about that. Again, as a dad, the AI seeping into... I mean, we usually talk about AI in the sense of the LLMs and chatbots and interactions that way, but toys is a whole other ballgame.
Joseph Gordon-Levitt: Yeah. Well, we’re not against our kids using technology. My kids do use computers quite a bit. I think there’s lots of great stuff you can do. Shout out to Beast Academy and The Art of Problem Solving. That’s the one we found that teaches math so super well. One of my sons is really into music and makes music with a computer, amazingly. And we let him spend a lot of time doing that. These kinds of tools are different than addictive algorithms that are designed to prey on your human weakness and suck up as much of your time and attention as possible and make sure that you keep coming back. That’s the same technique as a slot machine, really. There are reasons why there are laws that don’t let kids go into casinos and play slot machines. And the psychological techniques used to get gambling addicts hooked to slot machines are the same exact psychological techniques studied at Stanford and other places in Silicon Valley that make their products addictive.
Andrew Nusca: For sure.
Joseph Gordon-Levitt: It doesn’t make sense to have kids using these things. It’s bad for their brain. And then when you get into AI companions...
Andrew Nusca: Yeah, yeah, that’s even deeper.
Joseph Gordon-Levitt: When you’re a kid. As far as I understand, I was talking to the psychologist Jonathan Haidt about this, and any parent out there, I highly recommend his book The Anxious Generation. He was saying there’s quite a bit of scientific evidence that shows that when young brains are developing, it’s human interactions that really help develop those neural pathways. And an AI toy, or whatever chatbot you want, will imitate those human interactions, but they don’t develop those neural pathways in the same way, because it’s fake. Because you’re not talking to a human, you’re talking to an algorithm that’s designed to hook you and keep you and ultimately serve you ads. It’s very different than speaking to another person. And whether we’re talking neurologically, or you should just use your common sense as a parent, you’re trying to raise a kid who’s going to be a strong and compassionate and smart, feeling person.
Andrew Nusca: Indeed.
Joseph Gordon-Levitt: To me, it’s pretty obvious that you’re going down a very bad path if you’re subjecting them to this synthetic intimacy that these companies are selling.
Andrew Nusca: Who knew that the 90s chat rooms would be the good thing?
Joseph Gordon-Levitt: Yeah. Because it was other humans. I love those. Think of that.
Andrew Nusca: Think of that. Anyway. All right, let’s change gears a little bit and talk about you as a creator.
Joseph Gordon-Levitt: Okay.
Andrew Nusca: Because, I mean, you just said one of your kids plays music. I do too. There are AI music tools. How do you feel about that? You have definitely strong opinions, working in Hollywood, in terms of the impact of AI on creativity. The dangers are clear. Do you find any hope in what you can use as a person who acts, produces, writes?
Joseph Gordon-Levitt: A hundred percent.
Andrew Nusca: Creative goodness in all that?
Joseph Gordon-Levitt: For sure. The idea that, you know, like, I used to make so many videos when I was a little kid. And the idea that a couple of 10-year-olds or something could use these tools to make a movie and have it look as good as the Hollywood big shots can make their movies look, that’s great. That’s unequivocally great. My only concern is, where’s the economics? Because right now the way the companies are positioning it, they’ve built their models out of stolen content and data. Didn’t ask permission. They aren’t offering compensation. Or they are kind of offering compensation here and there, but mostly they’re claiming, “No, this is fair use, we don’t have to offer compensation,” which sort of turns them into hypocrites. They’ve paid a few companies, or maybe I guess they’re just covering their bets, or maybe it shows that they don’t actually have a moral stance here, they’re just trying to make money. But they’ve built their models out of stolen stuff. And their claim is that, okay, whenever these tools are used, 100 percent of the economic upside should go to the tech companies, and zero percent of the economic upside should go to all the people whose writing and voices and art and creativity and labor and time and skill and perspective went into making these products valuable. These products would have zero value, would be able to do zero things, without all of that human input. So why shouldn’t some of those humans get paid? I think it’s actually a principle that applies far beyond entertainment. This is something that, if we’re going to have a functioning economy moving into the future as this tech gets more and more common, we need to establish the principle that if someone does some good work, has a good idea, makes that idea happen, puts it out into the world, they’re allowed to gain economic value from it. But if an AI company is allowed to just take it for free and monetize it without paying that person, how is capitalism going to work anymore?
Andrew Nusca: Right, right.
Joseph Gordon-Levitt: So I see it going down a pretty dystopian road unless we can establish this principle that your digital self, your thoughts, your ideas, your work belongs to you, and any economic upside from that should be shared with you.
Andrew Nusca: Totally, totally. I want to come to the audience for one question before we wrap, but I want to ask you a very simple one before that, which is just: if some of these things were worked out, you, as a, I will just call you a Hollywood creative to take all the roles and wrap them into one, would you in fact use AI tools for your work?
Joseph Gordon-Levitt: I’m not against the technology.
Andrew Nusca: Yeah.
Joseph Gordon-Levitt: If it were set up ethically and people were getting fairly compensated, and we had all the proper upgrades to our systems to make this work fairly, absolutely. This could propel a great leap forward in creativity. But I think if we leave it as it is, it will only just serve to concentrate power in the hands of a few enormous companies, and that’s not really good for anybody.
