The Justice Department does not file briefs in private copyright disputes because it has nothing better to do. When the federal government intervenes in litigation between a tech company and a group of news organizations, it is making a policy choice — about who owns the fruits of journalism, about what the press is worth, and about which industries the state will protect. The brief the DOJ filed in support of OpenAI in a landmark copyright lawsuit brought by The Intercept and other media companies is that policy choice, made explicit.
The argument the government is advancing is deceptively simple: that training an AI model on copyrighted journalism without permission qualifies as fair use, and that news organizations therefore have no legal claim to compensation. But the simplicity is the tell. What looks like a statutory interpretation question is actually a structural decision about who gets to profit from the labor of reporters — and who does not.
The lawsuit is not a niche dispute. It is one of several cases working through federal courts that will determine whether the entire economic model of AI development rests on a foundation of uncompensated extraction. OpenAI, like its competitors, trained its large language models on enormous quantities of text scraped from the internet — including investigative journalism, longform analysis, and news reporting produced by organizations that paid reporters to gather, verify, and write it. The companies that built those models are now worth hundreds of billions of dollars. The newsrooms whose work fed those models are, in many cases, laying off staff or shutting down entirely.
The government's intervention is not legally required. The DOJ chose to file. That choice reflects an alignment of interests that predates this specific case: the current administration has been aggressively solicitous of the AI industry, working to strip away regulatory friction, preempt state-level AI laws, and now, apparently, insulate AI companies from the copyright claims of the press. As Tinsel News has documented, the DOJ has already sided with Elon Musk's xAI against a Colorado AI regulation — a move critics characterized not as legal reasoning but as industrial policy dressed in statutory language.
Follow the money and the pattern clarifies. OpenAI has cultivated relationships across the federal government with notable speed. The company secured a Pentagon contract while its CEO was publicly positioning OpenAI as a responsible actor in the AI space — a posture Tinsel News examined in its reporting on OpenAI's Pentagon deal and its simultaneous ethics theater. The AI industry has also deployed significant political spending to shape the Democratic Party's posture toward regulation, buying influence across both parties at a scale that has drawn comparisons to the most powerful lobbying operations in Washington.
The DOJ brief is the return on that investment.
The fair use argument that the government is advancing deserves scrutiny on its own terms, because it will not survive contact with what fair use was designed to do. The doctrine exists to protect transformative uses of copyrighted material — a critic quoting a passage, a scholar analyzing a text, a satirist parodying a song. It was not designed to immunize a for-profit corporation that ingests millions of copyrighted works wholesale, uses them to build a commercial product, and then sells that product in direct competition with the sources it consumed. OpenAI's models do not merely reference journalism. They reproduce its patterns, its analysis, its voice, and its facts — and then deliver those outputs to users who no longer need to visit the publications that produced the underlying work.
That is not transformation. That is substitution. And substitution at scale is what is killing the press.
Under U.S. copyright law, fair use is evaluated on four factors: the purpose and character of the use (commercial vs. educational), the nature of the copyrighted work, the amount used, and the effect on the market for the original. Legal scholars have argued that AI training fails on the fourth factor — market substitution — because AI outputs can replace demand for original journalism. The government's brief contests this framing.
There is a second dimension to this intervention that the framing of a copyright dispute tends to obscure: what it means for press freedom when the government takes sides against the press in a legal proceeding. The current administration has been conducting a broad campaign against media institutions it regards as hostile — threatening broadcast licenses, deploying the FCC as a political pressure tool, and subpoenaing journalists covering the Iran war. The DOJ's intervention in the OpenAI copyright case fits this pattern — not as a press freedom violation in the traditional sense, but as a structural weakening of the economic foundations that make independent journalism possible.
The Intercept, one of the plaintiffs in the lawsuit, is not a legacy institution with a century of institutional inertia. It was built as an independent, adversarial outlet — the kind of organization that covers stories inconvenient to powerful interests and relies on legal protections and reader support rather than corporate advertising relationships. When the government files a brief arguing that The Intercept's journalism can be ingested without compensation by a company worth more than most countries' GDP, it is not making a neutral legal argument. It is choosing a side in a fight about whether independent journalism can remain economically viable.
The broader copyright litigation involving AI companies and media organizations is not limited to The Intercept's case. The New York Times has filed a separate lawsuit against OpenAI and Microsoft. Other publishers have entered into licensing agreements — some voluntarily, some under legal pressure — that suggest even the AI industry recognizes, at some level, that the extraction model has limits. But those licensing deals, where they exist, were reached under the shadow of litigation, not because AI companies spontaneously decided to compensate the press. The DOJ brief, if it shapes the court's thinking, would remove that shadow entirely.
What the government is effectively arguing is that the AI industry should be allowed to build its products on the infrastructure of journalism without paying for it — and that this is not only legally permissible but good policy. The administration has not made that argument publicly in those terms, because stated plainly it is difficult to defend. Instead it hides inside fair use doctrine, inside statutory interpretation, inside the neutral-sounding language of amicus briefs. But the effect is the same: a wealth transfer from the organizations that produce journalism to the companies that profit from consuming it, blessed by the federal government.
Courts will ultimately decide whether the DOJ's legal theory holds. But the political signal the brief sends does not wait for a ruling. Every news organization watching this case — every editor calculating whether investigative journalism is worth the cost, every reporter wondering whether their work will be scraped and monetized by a company that pays them nothing — now knows where the government stands. Not with the press. With the platform.
