The US government filed an amicus brief last week in the ongoing copyright litigation against OpenAI, and it’s not subtle about where it stands.

“The United States has a strong interest in continuing to develop a robust and competitive artificial intelligence industry that sets the standard for the practice and procedure of AI use globally,” the brief reads.

Translation: We’re backing OpenAI. National competitiveness beats copyright law.

This isn’t just about one lawsuit. It’s the clearest signal yet that the federal government views AI development as a strategic priority that trumps intellectual property protections. And if you’re a writer, artist, or anyone who creates things for a living, you should pay attention to what just happened.

What the government is actually arguing

The Justice Department isn’t saying copyright doesn’t matter. They’re making a more sophisticated argument: that training AI models on copyrighted material without permission should count as fair use under existing law.

Fair use is the doctrine that lets you quote a book in a review or use a clip in a documentary without asking permission. It’s meant to balance the rights of creators against the public benefit of allowing certain uses of copyrighted work.

The government’s position is that vacuuming up millions of copyrighted works to train a language model falls under that same principle. It’s transformative use. It serves the public interest. It doesn’t substitute for the original work.

There’s a logic to this argument. When ChatGPT generates a response, it’s not reproducing the specific articles or books it trained on. It’s doing something different with that material. That’s the kind of transformation fair use is supposed to protect.

But here’s what makes this brief significant: the government isn’t leaving this to the courts to figure out on their own. They’re explicitly weighing in to say that America’s position in the global AI race depends on this interpretation winning.

The global competition angle

The brief’s emphasis on international competitiveness isn’t accidental. China isn’t wringing its hands about copyright when training AI models. Neither is the EU, really, despite their generally stricter approach to tech regulation. If the US courts decide that AI companies need to license every piece of training data, American labs operate under constraints their foreign competitors don’t face.

That’s a real concern. There’s a legitimate argument that hamstringing US AI development over copyright issues hands the advantage to countries with fewer scruples about intellectual property.

But let’s be clear about what’s happening here. The government is making a policy choice to prioritize AI development over creator compensation. They’re deciding that the economic and strategic benefits of powerful AI systems outweigh the economic interests of people whose work trains those systems.

You can agree or disagree with that choice, but you can’t pretend it’s not a choice. It’s an explicit decision about whose interests matter more.

What this means for creators

If this interpretation of fair use prevails, it creates a new normal for how AI relates to creative work. Your writing, your art, your code, all of it becomes fair game for training data. Not because you consented, not because you were compensated, but because the courts decided that building better AI models is more important than your control over how your work gets used.

The counterargument from the AI labs is that this is no different from how humans learn. Writers read other writers. Artists study other artists. Nobody asks permission for that, and nobody pays licensing fees.

But humans don’t learn by processing millions of examples and then competing directly with the people whose work they studied. When an AI trained on your writing produces content that substitutes for your services, that’s not the same relationship at all.

The government’s brief doesn’t really grapple with this. It treats the question as primarily about innovation and competition, not about the economic relationship between AI systems and the people whose work makes them possible.

The lobbying paid off

OpenAI and other AI labs have spent the last two years building relationships in Washington and making the case that AI development is a national priority. This brief is evidence that the strategy worked.

Microsoft’s restructuring announcement today, splitting its business into “Agents and Infrastructure” segments and finally disclosing Azure revenue separately, shows how completely AI has reoriented the tech industry’s biggest players. When Microsoft reorganizes its entire corporate structure around AI, and the government files briefs supporting the legal framework that makes that possible, you’re watching an industrial policy take shape in real time.

The question isn’t whether AI development is important. It obviously is. The question is whether we’re making considered decisions about the tradeoffs, or whether we’re just accepting that whatever is good for AI companies is good for America.

What happens next

This is an amicus brief, not a court decision. The judges in the OpenAI copyright cases will make their own determinations. But the government’s position matters. It signals to courts that there are broader policy implications at stake. It gives legal cover to rule in favor of the AI labs.

If the fair use argument wins, expect every AI company to treat all publicly available content as free training data. The battle will shift to new fronts: transparency about what data was used, compensation mechanisms outside the copyright system, maybe new forms of licensing.

If it loses, the AI industry faces a massive practical problem. You can’t put the genie back in the bottle, but you might be able to demand payment for future genies. That could mean compulsory licensing schemes, industry-wide settlements, or creative new approaches to data rights.

What won’t happen is the status quo. We’re past that. The government just made sure of it.

The real issue isn’t whether AI labs can legally train on copyrighted work. It’s whether we’re comfortable with a system where the people who create the raw material for AI get nothing, while the companies that process it into products capture all the value. The government’s brief suggests Washington has made its peace with that arrangement.

The rest of us are still figuring out if we should.