First Amendment questions at center of lawsuit against Minnesota's AI nudification ban

NINA MOINI: Another top story today. Elon Musk is taking Minnesota to court. The lawsuit by Musk's artificial intelligence company, xAI, argues the state's new so-called nudification ban, set to take effect Saturday, August 1, violates the First Amendment. The ban targets technology that allows users to create and post deepfake videos or images of people naked without their consent.

Companies that violate the law could face a hefty fine. It was passed almost unanimously by the legislature this year. Here to help us better understand the case and what this means for Minnesota's law is Jeff Kosseff. He's a legal expert on the First Amendment. Thanks for your time this afternoon, Jeff.

JEFF KOSSEFF: Thank you so much for having me.

NINA MOINI: Could you start by breaking down the case that is being made in this complaint and how it touches on the First Amendment? Most people here doing something like that to another person and know that that is wrong, but what is the First Amendment case?

JEFF KOSSEFF: Yeah, so AI raises some really fascinating legal issues. And one of the really interesting things that really has to be figured out is who has the First Amendment rights when you're using AI.

Is it the technology company? Is it the person who's prompting the AI? Is it the person receiving the information? And the First Amendment protects a lot of very controversial speech with the understanding that there could be a chilling effect.

So the case being made here is that this law that was passed by Minnesota that goes into effect on Saturday says that if someone uses your technology to nudify someone, which includes the artificial depiction of an intimate part, that that can lead to up to $500,000 in liability for the technology provider, regardless of whether the person who's the subject of the image consented or regardless of whether the technology provider knew of the use or even tried to stop the use.

So it's basically strict liability. It says if someone uses your technology to create this image, you can face a half million dollars in fines. So the issue is, does that create a chilling effect on the use of this entire technology for this particular goal, which is clearly a noble and worthy goal, to cut down on the use of non-consensual artificial images?

NINA MOINI: Yeah. Do you think that xAI has a strong case here? Are there other examples that you can point to maybe in other states?

JEFF KOSSEFF: So I think that there are a lot of really unsettled legal questions that I think will come out through this case, as well as other litigation in the near future. I think xAI does have strong arguments for a few reasons. First is that, as I said, it applies regardless of the technology provider's state of mind.

So even if it's saying we really want to take steps to cut down on the use of this technology for these purposes, they could still be liable. The definition of an intimate part, which is basically what triggers the liability, is very broad. It includes, for example, someone's inner thigh.

It applies to satire or parody, even of public officials. And really probably most importantly is it applies regardless of whether the individual who's the subject consents. And that really distinguishes it from a lot of the other laws that go after non-consensual images.

So in terms of other instances, this is a pretty unique law, again, because it's going after the technology provider, but we have seen the beginnings of some other legal disputes involving AI and deepfakes.

In California, for example, they had passed a law that restricted the distribution of materially deceptive election-related material, including AI-generated images of candidates. And last year, a court blocked much of that law on First Amendment grounds.

Now, Minnesota probably has a stronger case than California because in California, they were getting to really protected political speech. And it's probably much easier to defend a law that goes after AI-generated intimate images, so that might be an area where Minnesota might have a stronger case than California.

NINA MOINI: Yeah, and Minnesota Attorney General Keith Ellison posted on X yesterday that he's prepared to defend the law in light of this complaint. But what is going on with the relationship generally between the First Amendment and AI? How does it apply to AI? It seems like people are still trying to figure it out. Does it cover perhaps AI chatbots? Are users rights prioritized?

JEFF KOSSEFF: Well, so that really gets to what I find, as a legal scholar, a really fascinating issue, which, again, is whose rights are at stake here? So I think there's clearly the rights of the companies. So companies have First Amendment rights for expressive technology.

But I think the more interesting issue is what about the people who use AI and received AI-generated images? So the Supreme Court for decades has held that there's a First Amendment right to receive information, not just to create and express information.

So I think what all of these legal disputes are going to end up sorting out-- and I think it's probably going to take quite a few years until we have any clear guidance from the courts on this-- is when you're dealing with the regulation of AI, are we looking at the interests of the technology companies, of the people who prompt the technology, or of the people receiving the information?

My guess is it's probably going to be a combination of all of those things. But I think exactly where the limits come out and what is permissible regulation really has to be articulated by the courts in cases such as this.

NINA MOINI: Is there a difference between if the subject of this material is an adult or a child?

JEFF KOSSEFF: I think that clearly would be one of the differences because in First Amendment analysis, you look at the government's interests in regulation. In legal terminology, is there a compelling interest?

And any images of children, for example, have always been regulated when it comes to intimate images of children. There's not any debate about whether that could be regulated.

NINA MOINI: Children will still be protected is, I think, what people want to.

JEFF KOSSEFF: Yeah.

NINA MOINI: So this law was set to go into effect Saturday. What does it mean when a law doesn't go into effect and everything pauses? Does this mean technology companies can just stop what they were doing, or how long would it be on hold, do you think?

JEFF KOSSEFF: Well, so I'm not sure what the plan is in terms of how long it would be on hold.

My prediction will be that in the next week, we're going to see a lot of machinations in the court, perhaps for a request for a temporary restraining order or movement for a preliminary injunction, which would basically ask the court to at least make a preliminary ruling on the constitutionality before coming out with a final ruling.

So basically, to say, let's just freeze this in place for right now because what xAI is saying is that it can't offer the technology with this law in place because if it's facing liability of a half million dollars per image, regardless of what it's been doing or what it knows about, that that just really-- there's no business case to offer the technology, at least in Minnesota.

NINA MOINI: All right, we'll be keeping an eye on it. Thank you for your analysis, Jeff. We appreciate your time.

JEFF KOSSEFF: Thanks so much.

NINA MOINI: Jeff Kosseff is an incoming law professor at the University of Minnesota.