The US government has officially taken sides in the AI copyright wars, filing a legal brief that argues copyright restrictions could harm American AI competitiveness and national security. In an amicus curiae brief filed in the New York Times versus OpenAI lawsuit, Associate Attorney General Stanley Woodward Jr. contended that strict copyright enforcement would disadvantage US AI companies against foreign competitors.

What Did the Government Actually Argue?

The filing makes a striking claim: "Rules of law that make it significantly more difficult to develop a robust AI industry in the United States therefore threaten national security and give a competitive advantage to foreign adversaries who are not so encumbered," according to Woodward's brief. The government's position essentially frames AI training on copyrighted material as a national security imperative, not merely a business convenience.

The brief raises concerns that strict fair use rulings would create barriers to entry for smaller competitors. "An erroneous fair use ruling would hamper competition in the market for large language models, because only the largest technology companies might have the capital necessary to pay licensing fees," the filing states. This argument echoes claims made by OpenAI and other AI developers for years, but now carries the weight of official US government endorsement.

The government also contends that the transformative value of AI training outweighs market harm to copyright holders. "Any effect on the market for the copyrighted works cannot overcome the significant world-changing value of the training use's transformative purpose," the brief argues.

Why Are Copyright Holders Pushing Back So Hard?

The New York Times, which filed the original lawsuit, rejected the government's framing entirely. "The Administration is siding with a handful of trillion-dollar AI companies at the expense of the countless American creators whose work they stole," said Graham James, a New York Times spokesman. "Both AI and creators can thrive; AI companies simply need to pay fairly for the content that makes their products possible, as copyright law requires."

This dispute reflects a fundamental tension in copyright law. AI training typically involves downloading, copying, storing, and transforming copyrighted works into machine-readable formats like tokens and embeddings. Each of these steps potentially triggers copyright protections under existing law. The question is whether these activities fall within fair dealing exceptions, or whether they constitute infringement that requires licensing.

What Does Indian Law Tell Us About Fair Use and AI?

While the US government makes its case, courts in other countries are already grappling with these questions. India's Delhi High Court issued a significant interim ruling in July 2026 in the case of ANI Media Pvt Ltd v OpenAI Opco LLC, offering insights into how courts might analyze AI training under fair dealing frameworks.

The Delhi court declined to grant ANI an injunction against OpenAI and held that OpenAI's storage and use of ANI's news articles to train ChatGPT could potentially fall within India's fair dealing exception for private or personal use, including research. Importantly, the court did not automatically reject the defense simply because OpenAI is a commercial company. This suggests that commercial purpose alone may not disqualify AI training from fair dealing protection.

However, the Indian court also made a critical distinction: it separated the legality of the training process from the legality of the outputs. Training may involve internal processing of works without presenting them to the public, but outputs delivered to users could still infringe copyright if they reproduce substantial portions of protected works. This two-stage analysis prevents the legality of training from automatically protecting downstream copying.

How Are Courts and Lawmakers Weighing These Competing Interests?

The legal landscape remains unsettled. India's Copyright Act of 1957 contains a fair dealing framework but does not expressly create a comprehensive text-and-data-mining exception. Indian courts have traditionally assessed fair dealing through contextual inquiry, considering factors including the purpose of use, the amount and substantiality of material taken, the manner of use, the degree of transformation, and the effect on the copyright owner's market.

These factors are not applied mechanically. Instead, courts ask whether the defendant's use competes with or substitutes for the original work, and whether the use is consistent with the purpose of the statutory exception. This flexible approach contrasts with the US government's more categorical argument that national security and innovation should override copyright concerns.

• Training Process Legality: Whether copying, storing, and processing works for model training falls within fair dealing or research exceptions remains contested across jurisdictions.
• Output Reproduction: Even if training is lawful, outputs that reproduce substantial portions of copyrighted works may still constitute infringement.
• Commercial Purpose: Courts are divided on whether commercial operation automatically disqualifies AI training from fair dealing protection.
• Market Harm: The degree to which AI training harms the market for original copyrighted works is a key factor in determining fair use.
• Transformative Use: Whether AI training is sufficiently transformative to justify copying without permission remains a central legal question.

What Should Companies and Creators Know Right Now?

The US government's filing does not settle the law; it is merely an opinion the court can consider and reject. The New York Times case is still in early stages, with amicus briefs due by October 16, 2026. The case began in December 2023 and has not moved significantly since.

For AI companies, the government's support provides some political backing but no legal certainty. For creators and copyright holders, the filing signals that the Trump administration prioritizes AI competitiveness over copyright protection, but courts may reach different conclusions. The Indian court's interim ruling suggests that fair dealing analysis will be fact-specific, examining the nature of the training process, the storage of works, the allegations concerning outputs, and the evidence available.

The outcome of the New York Times case could reshape how AI companies approach training data globally. If courts reject the government's arguments and rule that copyright licensing is required, AI companies would face significant new costs. If courts accept fair dealing defenses, copyright holders would lose leverage to negotiate licensing agreements. Either way, the next few months will clarify whether AI innovation and copyright protection can coexist, or whether one must yield to the other.