Requires AI companies to report copyrighted works used to train their models to the government.
This bill would make companies that build generative AI models tell the government which copyrighted materials they used. They would have to send a detailed notice to the Register of Copyrights. This aims to create more transparency for copyright owners and the public about what content is used to develop AI.
Today, there is no federal rule requiring companies to tell the government which copyrighted works they use to train generative AI models. This bill would require companies to submit detailed notices to the Register of Copyrights, providing a sufficiently detailed summary of these works. It would also create a public online database of these notices. Copyright owners could sue companies that do not comply, with fines starting at $5,000 for each missed notice, up to $2,500,000 per year.
S 3813 · 119th Congress · February 10, 2026 · AI Summary by gemini-2.5-flash · 8/10
Sign in to see your representatives' phone numbers
12 filings mentioned this bill
Amounts reflect total quarterly lobbying spend reported to the Senate, not bill-specific spending. Source: Senate LDA filings.