The DOJ Just Chose a Side. Publishers Should Notice.

The Justice Department didn't file a neutral brief. It urged a federal court to treat the use of copyrighted works to train AI models as fair use, arguing that the public, competitive, and national security benefits of AI weigh heavily against imposing licensing liability.

For publishers, that's not a small distinction. It's the federal government taking a position on one of the most consequential questions facing the information economy.

In a statement of interest filed September 1st in re OpenAI, Inc. Copyright Infringement Litigation (MDL No. 25-md-03143), the DOJ told a Manhattan federal court that it was in the national interest for the judge to find that OpenAI did not violate copyright law when it used articles by The New York Times and other publishers to develop artificial intelligence systems. The department wrote that "the creative possibilities and public benefits" of training large language models on copyrighted content "far outweigh any competitive harm."

I've spent my career in ad tech partly because I believed the industry had a real, if imperfect, purpose: fund the journalism that keeps people informed. Revenue for publishers means investigative reporters stay employed. It means local newsrooms survive. It means the information ecosystem that a functioning democracy depends on doesn't collapse quietly.

Now the federal government is telling a federal judge that protecting that ecosystem is less important than protecting OpenAI's training pipeline.

That's the actual argument being made. It deserves a direct response.

The National Security Frame Cuts Both Ways

The DOJ argued that ruling for OpenAI will bolster American innovation and national security capabilities. That's a real concern. I'm not dismissing it.

But the same logic applies in the other direction, and the DOJ chose not to make that case.

An informed public is not a soft, feel-good concept. It is a hard prerequisite for a functioning democracy. Professional journalism, the kind that requires editors, lawyers, time, and accountability, is what produces the investigations that expose corruption, the reporting that shapes policy, and the context that prevents misinformation from winning by default. Strip the economic foundation from newsrooms and you don't get AI plus journalism. You get AI minus journalism.

The DOJ's brief argued that it "is not in the public's interest for the largest technology companies to have an oligopoly on LLM training due to licensing entry barriers" that function primarily as subsidies for legacy media. The framing there is worth sitting with. Paying for content you used to build a trillion-dollar business is now characterized as a subsidy for old media. That is a remarkable rhetorical move.

What This Means for Publishers

The Times has been joined by other news organizations and publishers in challenging how their work has been used to develop AI systems. This isn't simply a fight between one powerful media company and one powerful technology company.

It's a fight over whether publishers can require compensation when their copyrighted work is used as training material for commercial AI systems.

If courts ultimately adopt the DOJ's view of training-stage fair use, one of the most important sources of leverage publishers have in AI licensing negotiations could shrink considerably.

That does not mean publishers would have no leverage at all. Real-time access, APIs, retrieval, attribution, proprietary archives, output rights, brand permissions, and commercial partnerships can all remain valuable.

But if training on publisher content is broadly deemed fair use, a major piece of publishers' licensing leverage disappears. The danger is that publishers stop being treated as partners in the information economy and start being treated simply as data sources.

The False Choice

Nobody should have to choose between advancing AI and sustaining the journalism those models were built on. That framing risks making one side look like an obstacle to progress when the real question is how innovation and original content can remain economically sustainable at the same time.

The Times argued in its original December 2023 lawsuit that AI companies were free-riding on its investment in journalism by using that work to build substitutive products without permission or payment. That argument is not inherently anti-innovation. It is the same concern any business would raise if another company used its core product at scale to build a competing commercial system.

If AI companies derive enormous value from copyrighted content, compensation should not be treated as incompatible with innovation. The goal should be a workable market structure that encourages new technology without eliminating the economic incentives to create the information that technology depends on.

AI needs high quality information. High quality information needs an economic model that supports the people and institutions that create it.

The question shouldn't be whether America chooses AI or journalism. It should be how we build an AI economy in which creating original information remains economically worthwhile.

Because an informed public isn't separate from national security. It's part of it.

At Aditude, our business is built around helping publishers sustain the economics of creating valuable content. That's why we'll be watching this case closely.