YouTubers sue Snap over alleged AI training copyright use

A group of YouTubers has filed a lawsuit against Snap, alleging copyright infringement over AI training data.

The plaintiffs say Snap used their YouTube videos without permission to train AI features like the “Imagine Lens.”

The case was filed on January 23, in the US District Court for the Central District of California.

The creators claim Snap relied on datasets such as HD-VILA-100M, which were meant for research, not commercial use.

They also allege Snap bypassed YouTube’s terms and licensing rules.

The lawsuit seeks damages and an injunction, and is led by the creators of the H3H3 channel, alongside two smaller golf channels.

🔗 Source: TechCrunch

🧠 Food for thought

Implications, context, and why it matters.

It is alleged in the complaint that Snap’s “Imagine Lens” feature was trained on the research-only HD-VILA-100M dataset, and that claim would need to be proved in court. A connection would have to be established between the plaintiffs’ allegedly scraped YouTube videos and Snap’s AI features.
Even with new rules, this kind of proof may still be hard for creators. Under the EU’s upcoming AI Act, only high-level summaries of training data are required, rather than work-by-work disclosures, so rightsholders are left to prove specific copyright infringements in court 1.

For tech operators and investors, the growing pile of lawsuits against AI model providers points to demand for compliance help. The nonprofit Copyright Alliance has counted over 70 copyright infringement cases filed against AI companies.
That opens room for platforms that reduce legal and financial exposure for AI developers. Tools can verify training-data provenance (where training data came from and under what rights), handle copyright licenses, and support transparency rules like those in the EU AI Act 1.
Another useful add-on could include copyright indemnity insurance. Rules in one jurisdiction, like the EU, can raise lawsuit risk in others, such as the U.S. 1.