In the world of artificial intelligence (AI), the line between innovation and copyright infringement can be a blurry one. Recently, Thomson Reuters scored a major victory in a case that could shape the future of how AI companies use copyrighted content. The case centers on a legal AI startup, Ross Intelligence, which allegedly used Thomson Reuters’ Westlaw legal database to train its AI system without permission. The ruling has important implications, not just for tech companies, but also for industries like entertainment, where copyrighted content is often the basis for creative AI tools.
The Case: Thomson Reuters vs. Ross Intelligence
In 2020, Thomson Reuters, the parent company of Reuters News, filed a lawsuit against Ross Intelligence. The media and tech conglomerate claimed that Ross illegally copied content from its Westlaw database—one of the largest legal research platforms in the world. The content in question wasn’t just basic information, but editorial work, including case headnotes that summarize legal rulings, which are copyright-protected. Ross used this material to train its own AI model, which was designed to help legal professionals with research.
Thomson Reuters argued that Ross was using its content to create a competing product, essentially taking the work of its editorial team without permission. The legal issue at stake? Whether Ross’s use of this copyrighted material could be considered “fair use,” a defense commonly used by tech companies that argue their work falls under legal exceptions for research, education, or other purposes.
The Ruling: Fair Use Doesn’t Apply
The court’s ruling was a clear win for Thomson Reuters. U.S. Circuit Judge Stephanos Bibas ruled that Ross’s use of Westlaw’s copyrighted content was not fair use. While the fair use doctrine allows for the limited use of copyrighted material in certain cases (such as commentary, news reporting, or research), as determined by the standard four-factor fair use test, the safe harbor did not apply here because Ross was directly competing with Westlaw by using the copyrighted material to build a similar legal research tool.
Judge Bibas highlighted the fourth and most critical factor of the fair use test—the potential market harm—as the deciding factor. Ross’s AI model, he noted, undermined the market for Westlaw’s services, including future licensing opportunities for its database. In other words, Ross was essentially using Thomson Reuters’ work to take a slice of the market that had been built over years.
What Does This Mean for AI and Entertainment?
The ruling has wide-reaching consequences beyond the legal tech world. AI companies in the entertainment industry, such as those working with AI-generated scripts, music, and artwork, often train their systems on large datasets containing copyrighted works. As the AI industry grows, many creators, authors, and content owners are raising concerns about how their work is being used without their consent.
This ruling reinforces the notion that AI cannot freely use proprietary works to create competing products or services without proper licensing. For example, AI-driven music platforms that use existing songs to create new compositions or AI art generators that replicate visual styles may face similar legal challenges if they use copyrighted content without permission. As with the Thomson Reuters case, if these tools are seen as directly competing with existing markets, they might not be able to rely on the fair use defense.
Looking Ahead: What This Means for You
This case emphasizes the need for AI companies to carefully consider copyright law when building their tools. Whether you’re working with legal databases, visual art, music, or even writing, using copyrighted material without proper licensing could lead to costly legal battles.
This case also signals that courts are increasingly willing to hold AI developers accountable for willful copyright infringement, even if the work is used to fuel technological innovation. If you’re involved in any AI development, it’s crucial to consult with legal experts to ensure that your AI models are built in compliance with copyright laws.
As the AI industry continues to evolve, we can expect more cases like this one to clarify how copyrighted materials can be used in AI systems.
Reach out to our team to learn more about how we can support your business and help navigate the complex world of copyright and AI.
