Fine Print and AI: Examining Google’s Position on YouTube Training Data

As the music industry continues to fine-tune its approach to generative AI, the value attributed to training data is increasingly recognized, negotiated and realized through lawsuit settlements and emerging partnerships. While the central issue in most of these cases has been whether the platforms’ purported use of copyrighted materials constituted infringement or fair use, Google has surfaced a separate argument, seemingly simple, but far more complicated when taken in context with broader industry trends.

The Fine Print

In March 2026, a group of independent artists filed a complaint with the U.S. District Court for the Northern District of Illinois, claiming their songs were used to train Google’s AI music model, Lyria, without permission or compensation. Fundamentally, the artists’ argument is that they uploaded their music to YouTube with the intent of distributing it to their fans. But Google, being the parent company of YouTube, allegedly used that same music to train its AI model, which can now generate full-length tracks, complete with vocals and instrumentals, as part of the platform’s paid “Pro” subscription.

Although Google has not acknowledged the model was trained in this way, the company is seeking to dismiss the lawsuit altogether, on the grounds that such use would have been permitted through YouTube’s Terms of Service. According to the motion filed by Google in June 2026, when uploading their content to the website, the artists granted YouTube, and its “Affiliates,” a “worldwide, non-exclusive, royalty-free, sublicensable and transferable license” to “reproduce, distribute, [and] prepare derivative works” based on that content.

While Google’s position is firmly rooted in an explicit contractual arrangement between the parties, recent developments in the industry may have opened up the possibility to challenge the scope of the apparent license granted.

The Context

The high-profile settlements with Suno and Udio by Universal Music Group and Warner Music Group, as well as other partnerships and licensing deals within the industry, may reflect an emerging view that AI training data, especially when associated with copyrighted material, is not inherently royalty-free. In effect, these developments may suggest that training AI models on music is not seen as just another technical process, but a separate commercial use of creative works that requires distinct remuneration.

Accordingly, the courts could interpret these developments as evidence of market expectations. If judges see AI training as a fundamentally different kind of use, and not just another way to distribute or display content, they may be more hesitant to accept that a generic platform agreement would cover it.

On the other hand, if the courts choose to agree with Google’s position, it could solidify the norm that uploading content to a website, and agreeing to its standard Terms of Service, would allow such platforms to use that content for AI training without express authorization or payment.

The Big Picture

Ultimately, the YouTube lawsuit is not just about whether Google used artists’ works. It is about what artists actually agreed to when they uploaded them. Compared with Suno and Udio, the presence of a user agreement seems to give Google stronger legal footing. However, the trend toward licensing deals suggests that some stakeholders increasingly view AI training as a potentially compensable use. How courts reconcile contractual consent with these evolving economic norms will likely shape the next phase of AI governance and copyright protections in the music industry.

As these issues continue to evolve, artists and other rights holders are increasingly faced with complex questions around how their works are being used and whether they are appropriately participating in the value being created. By combining strong industry experience with deep technical expertise, Withum’s Royalty and Intellectual Property Audit Services can help clients gain greater visibility into their creative portfolio, recover underpaid royalties and make more informed decisions in light of changing market trends and economic conditions.