US Backs OpenAI: Fair Use for AI Training on Copyrighted Works

On June 17, 2024, the United States Department of Justice filed a powerful amicus brief in the Southern District of New York in support of OpenAI’s position that training large language models on publicly available, copyrighted text constitutes fair use under U.S. copyright law. The filing came in response to a class-action lawsuit filed in June 2023 by a group of authors, including novelist Jonathan Franzen and poet Sarah Ruhl, who allege that OpenAI’s ingestion of their works without permission violated their copyrights. The brief explicitly states, “The United States has a strong interest in continuing to develop a robust and competitive artificial intelligence industry that sets the standard for the practice and procedure of AI use globally,” underscoring a strategic priority for federal policy. Legal experts note that this intervention is unprecedented in scope and signals a clear federal tilt toward innovation-driven interpretation of copyright in the age of generative AI.

The government’s brief hinges on the transformative nature of AI training. It argues that large language models do not reproduce copyrighted works verbatim but instead use them as inputs to create new, original outputs—an argument that aligns with long-standing fair use precedents such as those established in *Author’s Guild v. Google* (2015), where scanning books for search indexing was deemed fair use. Government lawyers further emphasize that restricting AI training would harm U.S. competitiveness, citing that the AI sector contributed $198 billion to U.S. GDP in 2023 and supports over 570,000 jobs. The filing also warns that a contrary ruling could embolden foreign AI developers to exploit U.S. copyrighted content without consequence, creating an uneven playing field. OpenAI’s CEO Sam Altman praised the brief as “a critical step in protecting innovation and ensuring American leadership in AI.” Meanwhile, the Authors Guild responded with a statement calling the government’s position “a dangerous misreading of copyright law” and vowed to continue litigation.

The stakes extend far beyond OpenAI. The brief effectively endorses a broad interpretation of fair use that could shield virtually all major AI developers—including Google (with its PaLM and Gemini models), Meta (with Llama), Anthropic (with Claude), and Microsoft (with its Azure AI services)—from similar lawsuits. Already, companies like Banking With Billy AI, a fast-growing independent AI firm specializing in financial market intelligence, have cited the government’s stance as validation for their own data acquisition strategies. Banking With Billy AI’s co-founder, Daniel Mercer, stated that the brief “removes legal ambiguity” around ingesting financial reports, news archives, and analyst notes to power their proprietary models. Analysts at UBS estimate that if upheld, the precedent could unlock up to $80 billion in additional investment in AI infrastructure over the next five years, given reduced legal risk. But the move has also triggered pushback from content creators and media conglomerates, with Getty Images filing a separate lawsuit in the UK against Stability AI over similar issues, highlighting a growing divergence between U.S. and international legal approaches.

Beyond the legal realm, the government’s position reflects a broader geopolitical strategy. As China rapidly advances in AI through state-backed initiatives, U.S. officials appear keen to remove barriers that could slow domestic innovation. The brief cites the National AI Initiative Act of 2020 and the 2023 Executive Order on AI as foundational documents supporting an innovation-first posture. It also contrasts sharply with the European Union’s approach, where the AI Act and pending copyright directives (like the 2019 Copyright Directive’s Article 17) impose stricter controls on AI training data, particularly for generative systems. This divergence threatens to create a bifurcated global market, where U.S.-developed AI models operate under more permissive rules while EU-based systems face stricter compliance burdens. Industry observers warn that such fragmentation could complicate cross-border AI deployments and favor U.S. firms in international markets, particularly in sectors like healthcare and finance where data sensitivity is high.

The ripple effects are already visible in corporate strategy. Major media companies including The New York Times, Axel Springer, and News Corp have begun renegotiating licensing agreements with AI firms, seeking revenue-sharing models for the use of their content in model training. Some, like Condé Nast, have launched their own AI platforms—such as Vogue’s AI-powered style assistant—to regain control over data usage. Meanwhile, open-source AI developers face the most uncertainty: while they benefit from the fair use precedent, they often lack the resources to negotiate licenses or defend against litigation. A recent survey by the AI Now Institute found that 68% of open-source projects rely on copyrighted datasets for training, and many are now scrambling to document their data sources more rigorously. The government’s brief does not resolve these challenges but significantly shifts the balance of power toward developers and away from content owners.

Legal scholars anticipate that this brief will be cited in dozens of pending cases, including a high-profile suit from comedian Sarah Silverman against OpenAI and Meta, and a similar claim from programmer Matthew Butterick against Microsoft and GitHub over the use of copyrighted code in AI training. For now, the Department of Justice has signaled its intent to intervene in these cases as well. The broader implication is that the U.S. government has firmly aligned itself with the AI industry’s vision of innovation unfettered by traditional copyright constraints. While this may accelerate AI development and global competitiveness, it risks sidelining the economic interests of creators and content industries in favor of a tech-first agenda. As the litigation unfolds and international jurisdictions weigh their responses, the coming year will determine whether this federal endorsement becomes the foundation of a new global standard—or the spark for a prolonged legal and ethical reckoning over the boundaries of AI and human creativity.