The winners here are the labs with the biggest unlicensed-corpus exposure — OpenAI, Microsoft, and by extension Anthropic, Meta, and Google, all defendants in overlapping copyright suits. The losers are publishers, authors, and music labels who spent the last two years building licensing markets — NYT-Amazon, Axel Springer-OpenAI, Reddit-Google — on the assumption that training without a deal carried real legal risk. A Justice Department brief arguing that risk mostly doesn’t exist is a direct hit on that market’s pricing power.
What makes this notable isn’t the legal theory, which tracks arguments AI companies have made all along, but the source. According to NOTUS, DOJ attorneys have historically leaned toward copyright holders — citing Obama-era solicitor general Donald Verrilli’s work for MGM in the Grokster case — making this filing, per Re:Create’s Brandon Butler, a break from institutional norm rather than a routine amicus gesture. Copyright lawyer Evan Brown told Wired the presiding judge, Sidney Stein, isn’t bound by the brief but will “take it quite seriously” precisely because it carries DOJ’s institutional weight, and other judges hearing similar suits are likely to see it too.
If courts adopt the government’s framing, the legal necessity that justifies paying for training data evaporates — leaving voluntary licensing to survive only where publishers can sell something litigation can’t force: curation, freshness, and access.
That’s the real stakes for data markets. Licensing intermediaries positioning themselves as compliance infrastructure — content-licensing platforms built for a world where unlicensed training was legally dangerous — lose their strongest sales argument if fair use is settled law. The Anthropic case is the cautionary counterexample worth remembering: that $1.5 billion settlement, the largest copyright payout on record, was for pirating books, not for training on them — the judge held training itself was fair use. DOJ’s brief pushes every pending case, including new Sony and Warner Music suits, toward that same bifurcation: training stays protected, only sourcing method creates liability. Watch whether the DOJ files similar statements in the Anthropic-adjacent author suits and the Sony/Warner Music cases — that would confirm this is administration policy, not a one-off intervention, and would reprice every AI-content licensing deal signed on the assumption that courts might go the other way.
AI dominance is critical to promote national security, prosperity, and economic mobility for all Americans. This Administration will never let our Nation be at a disadvantage relative to our foreign adversaries based on a plainly incorrect understanding of copyright law.
