DOJ sides with OpenAI in AI copyright dispute with publishers

Facebook
Twitter
LinkedIn
Pinterest
Pocket
WhatsApp

DOJ sides with OpenAI in a closely watched copyright fight, backing arguments that training artificial intelligence systems on news articles can qualify as fair use and should not require licenses or payment to rights holders.

Several media organizations, including The New York Times, have sued OpenAI and Microsoft, alleging the companies used millions of articles to train AI models in violation of copyright. The defendants argue that using those materials to develop models is protected as fair use, a longstanding doctrine that permits certain uses of copyrighted works without permission or compensation.

The Department of Justice filed a statement of interest supporting OpenAI’s position, according to court filings. The filing advances a broad view of fair use that would allow AI developers wide latitude to train on copyrighted content for commercial purposes without securing licenses. Critics say the stance departs from what had been a more balanced posture within the administration and could tilt the legal landscape toward the largest technology firms.

DOJ sides with OpenAI in fair use debate

In its filing, the department argues that requiring licenses for training datasets would raise significant costs for AI developers. It says those costs could limit competition, disadvantage smaller firms, and undermine U.S. leadership and national security. The brief’s emphasis on innovation and competitiveness contrasts with other administration guidance that acknowledges unresolved questions at the intersection of AI and copyright.

Administration materials, including the White House’s AI framework, have noted that both creators and technology companies have legitimate interests in the debate. Similarly, principles adopted by G20 ministers at a recent Commerce Department innovation meeting described the AI–copyright nexus as raising complex legal and policy issues and underscored the essential role of intellectual property protections.

Startups and licensing advocates have countered that paying for content can provide more reliable access to training data, reduce legal risk, and improve model quality. Companies have also begun aggregating rights to streamline licensing. Observers contend that large AI companies are best positioned to pay for access and that doing so could help develop the very licensing markets they say are difficult to navigate.

Opponents of the DOJ’s approach argue that national security concerns do not permit government to disregard property rights. They point to constitutional and statutory requirements for just compensation when private property, including intellectual property, is taken for public use. Allowing training on copyrighted works without permission or payment, they say, would be akin to authorizing use of private assets without fair value.

If courts adopt the department’s expansive view, critics warn it could depress incentives to create books, journalism, research, films, and music. Copyright, they note, supports livelihoods for creators and underpins significant economic activity. Eroding those protections could also diminish the quality and quantity of future training data, ultimately harming AI development itself.

Commentators urge the administration to let courts resolve disputes case by case rather than pressing for sweeping fair use exceptions. They argue that targeted legislative steps would better balance innovation with creators’ rights.

Disclosure proposal aims to increase transparency

Lawmakers have introduced the bipartisan CLEAR Act, sponsored by Sen. Adam Schiff of California and Sen. John Curtis of Utah. The bill would require AI developers to disclose the copyrighted works included in training datasets for new generative models. While it would not determine whether particular uses qualify as fair use, the measure would give creators information to assess whether their works were used and to pursue compensation or remedies if they believe their rights were violated.

Proponents say such transparency could foster a functioning market for licenses rather than treating licensing as an impediment to innovation. They argue that strong AI progress and robust copyright can coexist if policymakers encourage efficient licensing mechanisms and respect for intellectual property.

Supporters of the DOJ’s position maintain that flexible fair use is essential to keep the United States competitive in AI. Opponents counter that sustainable innovation depends on preserving the economic incentives that drive creative industries. Both sides agree that AI’s future will be shaped by how courts and Congress resolve these questions.

Observers note that the debate echoes broader questions about how emerging technologies should be regulated, similar to past disputes over nuclear oversight involving the International Atomic Energy Agency.

Facebook
Twitter
LinkedIn
Pinterest
Pocket
WhatsApp