DOJ’s AI Fair Use Filing Sparks Debate Over Intellectual Property

The Justice Department's recent support for AI companies in copyright disputes faces scrutiny from legal and innovation experts who advocate for balanced protections.
A significant legal confrontation over artificial intelligence is testing the boundaries of “fair use” in copyright law, drawing attention to a shift in position by the Department of Justice (DOJ). As tech companies face lawsuits from news organizations, including The New York Times, over the use of protected material in training AI models, the DOJ has filed a statement of interest that largely backs the position held by OpenAI and Microsoft.
At the core of the dispute is the principle of fair use. AI developers argue that using massive datasets of newspaper articles and other copyrighted works to train generative AI systems qualifies as fair use, exempting them from the need for permissions or compensation. The DOJ’s recent filing adopts this view, arguing that mandating licenses for training material would impose significant costs on developers. The government suggests that these financial burdens could hinder smaller companies, ultimately stifling U.S. competitiveness in AI and affecting national security.
However, this stance is being met with criticism from legal experts, including Coke Morgan Stewart, president and CEO of the Council for Innovation Promotion. Stewart, who previously served as the deputy under secretary of Commerce for intellectual property, argues that the DOJ’s filing represents an “extreme vision” of fair use. She contends that this expansive interpretation threatens to turn a narrow exception into a blanket exemption for the world’s most profitable industries, effectively allowing them to bypass the established property rights of creators.
Critics of the DOJ’s position suggest that it departs from more measured guidelines elsewhere in the federal government. The White House’s official AI framework has previously acknowledged the existence of valid arguments on both sides of the copyright debate. Similarly, the Carolina Principles for Emerging Technologies, adopted by G20 ministers at a recent Innovation Ministerial, emphasized the “essential role” of copyright protections and identified the intersection of AI and intellectual property as a complex policy challenge that requires respect for creative works.
There is also disagreement regarding the economic implications of licensing. While the DOJ raises concerns about costs, some industry participants argue that paying for content could actually create a more stable environment. They note that purchasing access to training material reduces legal uncertainty, improves model quality, and encourages the emergence of new businesses designed to aggregate and license copyrighted works. Furthermore, skeptics of the DOJ’s argument point out that major AI firms are well-positioned to manage these licensing costs, which would help develop a functioning market for intellectual property.
According to a Fox News poll, public opinion reflects a sense of urgency regarding AI regulation, with voters prioritizing safeguards alongside innovation. This concern aligns with the argument that allowing AI companies to appropriate intellectual property without compensation could diminish incentives for creators. If the legal system permits broad fair use, there is a risk that the long-term supply of human-created works—books, music, and journalism—could decline, paradoxically reducing the quality of material available for future AI training.
Legislative efforts are already underway to address these transparency gaps. The bipartisan CLEAR Act, introduced in February by Sen. Adam Schiff (D-Calif.) and Sen. John Curtis (R-Utah), proposes requiring AI developers to publicly disclose the copyrighted works utilized in their training datasets. While the act does not resolve the underlying fair use debate, it aims to provide creators with the information needed to pursue compensation or remedies.
As the court system considers the arguments, the fundamental tension remains: balancing the drive for technological advancement with the protection of property rights. For proponents of stronger copyright enforcement, the path forward involves fostering a marketplace where AI companies can efficiently secure access to materials while upholding the rights of those who produce them.