Seattle Times and Newsday Sue OpenAI and Microsoft Over AI Training Data

The Seattle Times and Newsday have filed a new copyright lawsuit against OpenAI and Microsoft, escalating the rapidly expanding legal battle over whether artificial intelligence companies can use news organizations’ work to train and operate AI systems without permission.

The lawsuit, filed in federal court in New York on September 4, 2026, accuses OpenAI and Microsoft of using newspaper content without authorization or compensation. The case arrives only days after the U.S. government entered another major AI copyright dispute on OpenAI’s side, arguing that using copyrighted material to train large language models can qualify as fair use.

Together, the developments demonstrate how quickly AI copyright disputes are becoming one of the technology industry’s most consequential legal issues. The outcome could influence how AI companies acquire training data, how publishers negotiate licensing agreements and whether creators can demand payment when their work becomes part of commercial AI systems.

What the New Lawsuit Claims

The Seattle Times and Newsday allege that OpenAI and Microsoft used their journalistic work to train artificial intelligence systems without obtaining permission. The publishers argue that their articles are valuable copyrighted works and that AI companies should not be able to use those works as commercial training material without authorization.

The allegations fit into a much broader wave of litigation. Publishers, authors, artists, music companies and other rights holders have increasingly challenged technology companies over the collection and use of copyrighted material for AI training.

OpenAI and Microsoft have generally argued that training AI models is legally different from simply republishing the original works. Their position is that the models learn statistical patterns from large quantities of information and use those patterns to generate new outputs rather than storing and distributing the original material as a conventional database.

The publishers disagree, arguing that the training process can exploit the economic value of their work and that AI products can ultimately compete with the organizations that created the underlying content.

Why News Publishers Are Fighting Back

News organizations face a particularly difficult problem because generative AI can answer questions directly using information that originated in journalism.

Traditional search engines generally send users to publishers’ websites. AI assistants can instead provide a synthesized answer inside the AI interface. That difference has created concern among publishers that AI systems could reduce traffic, subscriptions and advertising opportunities even when the AI-generated answer is based on professionally reported journalism.

The issue becomes even more complicated when an AI system reproduces passages from an article or provides information so detailed that a user no longer needs to visit the original source.

Publishers therefore argue that the economic relationship between AI companies and journalism should involve licensing or other forms of compensation.

Some media organizations have pursued commercial licensing arrangements with AI companies, while others have chosen litigation. The growing number of approaches reflects the uncertainty surrounding how copyright law will apply to generative AI.

The U.S. Government Has Just Taken OpenAI’s Side

The new lawsuit comes at a particularly significant moment because the U.S. government recently submitted a brief supporting OpenAI in its separate copyright dispute with The New York Times and other publishers.

The Justice Department argued that AI training generally qualifies as fair use and emphasized the importance of maintaining U.S. leadership in artificial intelligence.

The government’s position is not a final judicial ruling. The court must still evaluate the arguments and evidence presented by the parties. However, the intervention demonstrates that AI copyright disputes are no longer simply private disagreements between technology companies and publishers.

The issue has become connected to broader questions about American technological competitiveness, innovation and national security.

Microsoft Is Making Its Own Fair-Use Argument

Microsoft has separately asked a federal judge to reject copyright claims brought by publishers and other rights holders.

The company argues that training large language models on copyrighted material is a transformative use and that the resulting AI systems serve a different purpose from the original works.

Microsoft has also pointed to analysis of millions of Copilot chat logs. According to reporting on the company’s court filings, Microsoft argues that substantial reproduction of copyrighted material in ordinary AI interactions is relatively uncommon.

That argument is important because copyright cases can involve two separate questions: how copyrighted material was used to create the AI model and whether the resulting AI product reproduces protected expression.

A company could argue that the training process is lawful even if certain outputs later raise separate copyright concerns.

The Battle Is Bigger Than OpenAI

Although OpenAI and Microsoft are at the center of this latest dispute, the legal uncertainty extends throughout the AI industry.

Anthropic, Google, Meta and other technology companies have also faced copyright lawsuits related to the use of books, journalism, images, music and other creative works in AI development.

That means the eventual court decisions could have consequences far beyond a single company.

If courts broadly accept the argument that AI training is fair use, developers may have greater freedom to train models on publicly accessible material. If courts instead determine that commercial AI training requires permission or licensing in many circumstances, the cost and structure of AI development could change significantly.

Why the Training Question Is So Difficult

Copyright law was developed long before modern machine learning systems existed. Courts therefore have to apply established principles to a technology that creates a fundamentally different relationship between source material and output.

One central question is whether training an AI model on copyrighted material is sufficiently transformative.

Another is whether AI systems harm the market for the original works.

Publishers argue that AI assistants can become substitutes for traditional news consumption. Technology companies counter that AI systems provide a new tool for discovering and transforming information rather than simply replacing the original publication.

Courts will also have to consider whether facts and ideas contained in copyrighted articles receive the same protection as the original wording used to express those facts.

These distinctions could become crucial as AI systems become better at summarizing current events and answering questions about specialized subjects.

AI Search Could Make the Dispute Even More Important

The copyright debate is occurring at the same time that AI is changing how people discover information online.

Traditional search results depend heavily on users clicking through to individual websites. AI search systems increasingly attempt to answer questions directly, sometimes using information from multiple sources.

For publishers, that creates a difficult economic question: if an AI system can provide the essential information from an article without sending the reader to the publisher’s website, who should capture the value?

The answer could shape the future business model of digital journalism.

If licensing becomes the standard, AI companies could end up paying publishers for access to high-quality news archives and real-time reporting. If courts instead establish broad fair-use rights, publishers may need to find other ways to maintain revenue as AI becomes a larger gateway to information.

What This Means for AI Companies

The lawsuits create financial and strategic uncertainty for AI developers.

Training frontier AI systems already requires enormous investments in computing infrastructure, engineering talent and data processing. If companies must also negotiate licenses for large quantities of copyrighted material, the cost of building advanced models could rise substantially.

At the same time, licensing agreements could provide a clearer legal foundation for AI development.

Instead of relying on unresolved fair-use arguments, companies could negotiate directly with publishers and other rights holders. This approach could also give creators a predictable revenue stream from the growth of generative AI.

The challenge is determining what constitutes a reasonable price and whether smaller AI companies can afford the same licensing arrangements as the largest technology firms.

What This Means for Publishers

For news organizations, the stakes are equally high.

Professional journalism requires reporters, editors, photographers, researchers and other staff. Publishers argue that AI companies benefit from this investment when their models learn from large collections of journalistic work.

At the same time, publishers recognize that AI tools could potentially expand their audience and provide new distribution opportunities.

The industry is therefore divided between litigation, licensing and experimentation with AI partnerships.

The legal decisions emerging from these disputes could determine which strategy becomes economically sustainable over the long term.

A Global AI Copyright Problem

The issue is not limited to the United States.

AI companies operate globally, while copyright laws differ between countries. Rules governing text and data mining, licensing and fair use can vary substantially across jurisdictions.

A model trained using information collected from websites around the world may therefore face different legal questions depending on where the data was obtained, where the model was trained and where the AI product is offered.

That complexity could encourage governments to develop new AI-specific copyright rules rather than leaving every dispute to traditional litigation.

What Could Happen Next?

The immediate future of the Seattle Times and Newsday lawsuit will depend on the court process, including motions, evidence and legal arguments over the use of copyrighted material.

Meanwhile, other major copyright cases are moving forward. Courts are increasingly being asked to determine whether AI training is transformative, whether AI outputs substitute for protected works and what obligations technology companies have toward creators.

The answers could eventually establish a new legal framework for the AI economy.

For technology companies, the goal is to preserve the ability to train powerful models at scale. For publishers and creators, the goal is to ensure that AI development does not turn copyrighted work into uncompensated commercial infrastructure.

Both sides are likely to continue pushing their arguments until courts provide clearer guidance.

Conclusion

The Seattle Times and Newsday lawsuit adds another major front to the rapidly expanding AI copyright battle.

The case is important not simply because two newspapers are challenging OpenAI and Microsoft, but because it comes during a broader confrontation over who owns the economic value generated from AI training data.

OpenAI and Microsoft argue that training models on copyrighted material can be a transformative use protected by fair-use principles. Publishers argue that their journalism is being used as valuable commercial input without permission and that AI products could ultimately compete with the businesses that produced the content.

There is no simple answer yet. Courts will have to apply existing copyright principles to a technology that has fundamentally changed how information can be processed and delivered.

Whatever the eventual rulings, the consequences will reach far beyond newspapers. They could influence the economics of AI development, the future of digital media and the relationship between technology companies and virtually every industry that produces original content.

FAQ

Why are Seattle Times and Newsday suing OpenAI?

The newspapers allege that OpenAI used their copyrighted journalistic content to train and operate AI systems without authorization or compensation.

Why is Microsoft also being sued?

The lawsuit names Microsoft alongside OpenAI. Microsoft is a major OpenAI partner and has incorporated OpenAI technology into products such as Copilot. Microsoft is also separately defending its own AI-related copyright claims.

Does AI training automatically violate copyright?

No. Whether AI training constitutes copyright infringement is an unresolved legal question in many jurisdictions. Technology companies generally argue that training is transformative and can qualify as fair use, while copyright owners dispute that interpretation.

Why does this matter for AI users?

The outcome could influence how AI companies obtain training data, how much they pay for licensed content and how AI assistants are allowed to use and reproduce information from copyrighted sources.

Could publishers receive licensing payments from AI companies?

Yes. Licensing is already one potential approach to resolving the dispute. Some publishers have negotiated commercial relationships with AI companies, while others are pursuing legal action.

When will the courts decide?

There is no final timetable for resolving all of the related AI copyright disputes. Several cases are moving through U.S. courts, and different decisions could eventually be appealed.

Sources

Reuters — reporting on the September 2026 Seattle Times and Newsday lawsuit against OpenAI and Microsoft.
Reuters — reporting on the U.S. government’s support for OpenAI’s fair-use position.
The Verge — reporting on Microsoft’s arguments and Copilot output analysis in the publisher litigation.
Associated Press — coverage of the broader AI copyright dispute between technology companies and publishers.

Seattle Times and Newsday Sue OpenAI and Microsoft Over AI Training Data

Seattle Times and Newsday Sue OpenAI and Microsoft Over AI Training Data

What the New Lawsuit Claims

Why News Publishers Are Fighting Back

The U.S. Government Has Just Taken OpenAI’s Side

Microsoft Is Making Its Own Fair-Use Argument

The Battle Is Bigger Than OpenAI

Why the Training Question Is So Difficult

AI Search Could Make the Dispute Even More Important

What This Means for AI Companies

What This Means for Publishers

A Global AI Copyright Problem

What Could Happen Next?

Conclusion

FAQ