Imagine waking up to find a fake nude image of yourself circulating online, generated by AI. Now imagine the government promising to ‘ban’ the technology that created it. You’d feel safer, right? Wrong.
You’ve probably heard about the new law in Minnesota—the first in the nation to try to outlaw AI ‘nudification’ tools. Sounds like a no-brainer, doesn’t it? Protect victims, punish the creeps. But here’s the uncomfortable truth you’re not being told: that law is a paper tiger. It can’t work. And Elon Musk’s xAI is suing not because they want to defend nudification, but because they’re the only ones willing to call out the emperor’s new clothes.
The law, signed in May and set to take effect this Saturday, bans the use of AI to create fake nude images of real people. Noble goal. But the problem is technical: the technology these laws target is built on open-weight AI models—models that anyone can download, modify, and run on their own hardware. You can’t ban a math problem. Open-weight AI models are just equations—you can’t make them illegal.
As one developer told me: ‘You can’t un-publish a GitHub repo. The cat is out of the bag.’ The state of Minnesota has no ability to police every laptop, every server, every cloud instance that runs these models. The ban is purely symbolic—a feel-good headline that does nothing to stop the actual abuse.
And that’s where xAI’s lawsuit comes in. On the surface, it looks like a tech billionaire defending a dystopian tool. But dig deeper. The real fight is about liability. xAI is arguing that the law tries to hold model creators responsible for how users choose to deploy their software. That’s a massive legal shift. If the state wins, every AI developer—from OpenAI to a solo coder in a basement—could be sued for downstream misuse. The real question isn’t whether we should protect people from deepfakes. It’s whether we’re willing to sacrifice the open internet for a false sense of security.
This is the tension nobody wants to talk about: the visceral fear of being victimized by AI imagery clashes with the anxiety of living in an over-regulated, restricted digital space. We want safety, but we also want freedom. The Minnesota law pretends we can have both by banning a tool. It can’t.
So here’s the choice: continue chasing symbolic bans that do nothing but create a legal minefield, or have an honest conversation about where liability actually belongs—on the people who misuse the technology, not the tools themselves. The Minnesota law is a well-intentioned failure. xAI’s lawsuit might be the uncomfortable push we need to finally get this right.
FAQ
Q: Isn't it obvious that we should ban AI nudification tools? What's wrong with that law?
A: The law is well-intentioned but technically unenforceable. Open-weight models can be downloaded and run anywhere, beyond state jurisdiction. The ban only creates a false sense of security while doing nothing to stop actual abuse. Worse, it sets a precedent for regulating open-source AI that could stifle innovation.
Q: If the ban doesn't work, what should be done instead?
A: Focus on punishing the behavior (non-consensual distribution) rather than the tool. Strengthen laws against harassment and extortion. Invest in detection and watermarking. And hold platforms accountable for hosting abusive content, not the model creators.
Q: Isn't xAI just trying to protect its business? Why should we trust them?
A: xAI's lawsuit is self-interested, but that doesn't make it wrong. They're arguing that model creators shouldn't be liable for user misuse – a principle that protects all open-source AI. If the state wins, it could set a dangerous precedent where any AI developer is responsible for what their users do, effectively killing open-source AI.