Rep. Adam Schiff introduced a bill that would require companies training generative AI models to disclose all of the copyrighted work they used to do so
including that it's propelled by a calculation that forcing AI companies to disclose copyrighted materials is the politically popular move https://www.billboard.com/... Luiza Jarovsky / @luizajarovsky : 🚨BREAKING: @RepAdamSchiff proposes a bill to require transparency from companies on their use of copyrighted material to train AI systems. What you need to know: ➡️The proposed bill - the Generative AI Copyright Disclosure Act - would require prior notice to the Register of... [image] Ed Newton-Rex / @ednewtonrex : Today the Generative AI Copyright Disclosure Act was introduced by @RepAdamSchiff, and it's a great step towards fairer data practices in gen AI. - AI companies will have to disclose to the copyright office a full list of copyrighted works used to train their models - Disclosure... Jess Miers / @jess_miers : So, compliance with this law will be expensive as hell for AI startups, many of which won't even know if they have a successful product to begin with before investing in training. But maybe lawmakers think that's a feature not a bug? [image] Jess Miers / @jess_miers : Copyright attaches automatically to any creative works fixed in a tangible medium of expression. So, pretty much all works used to train an AI system will require disclosures, regardless of fair use considerations. (btw you don't “train” a dataset but details). [image] Jess Miers / @jess_miers : BUT THAT'S NOT ALL! Datasets are incredibly dynamic, especially when it comes to AI training. So, each time the set is updated in a “significant manner,” the notice requirement is triggered. [image] @iatse : IATSE commends Rep. Adam Schiff for introducing the Generative AI Copyright Disclosure Act. Entertainment workers must be fairly compensated when their work is used to train, develop or generate new works by AI systems. This legislation will ensure there is appropriate... Thomas / @t_nyc : This half-assed $5000 cost-of-doing-business “penalty” to our billionaire employers is being sold by our @IATSE @WGAEast @sagaftra “leaders” as protection from AI. All so Adam Schiff, a billionaire owned corporate Dem who supports genocide, can appear “pro-Union” Nick Gillespie / @nickgillespie : Is there anyone more anti-progress than @AdamSchiff? https://thehill.com/... Adam Schiff / @repadamschiff : Yesterday I announced my Generative AI Copyright Disclosure Act. We've seen an outpouring of support - from creators, companies, and more. Thank you to @sagaftra, @directorsguild, @WGAWest, @ASCAP, @NAVAVOICES, and so many more who are supporting this critical legislation. Adam Schiff / @repadamschiff : While new technologies like AI offer boundless opportunity — we must also ensure creators and their copyrighted work are protected. [video] Todd O'Boyle / @ttoboyle : Rep @AdamSchiff just introduced a bill to regulate generative AI and it, I kid you not, covers spreadsheets and spellcheck. [image] @sagaftra : “Everything generated by AI ultimately originates from a human creative source. That's why human creative content—intellectual property—must be protected.” - SAG-AFTRA NED & Chief Negotiator @DuncanCI SAG-AFTRA is proud to support this legislation. @future_of_music : This is an interesting approach to AI legislation. Schiff's bill forces Generative AI developers to disclose what data is used for training, sidestepping the question of when use needs to be licensed. It poses the question: Don't we have a right to know? @wgawest : Generative AI can generate only because of copyrighted works—works these models have co-opted without consent, compensation, or transparency. WGAW is pleased to support @RepAdamSchiff's Generative AI Copyright Disclosure Act as a first step towards addressing this theft-based AI... @wgaeast : “Given the scope and potential threat of AI, enforceable regulations are urgently needed to keep companies from implementing this technology in the shadows, without people's consent or knowledge,” said WGAE president @LisaCullen Adam Schiff / @repadamschiff : AI has the potential to change our economy, our political system, and our day-to-day life in dramatic and potentially disruptive ways. My Generative AI Copyright Disclosure Act will further innovation by safeguarding the contributions of creators, ensuring that they are aware of when their work contributes to AI datasets. LinkedIn: Drew Sheridan : If passed, AI companies would need to provide the Copyright Office with a full list of copyrighted works in their model's training data within 30 days. … Forums: Msmash / Slashdot : New Bill Would Force AI Companies To Reveal Use of Copyrighted Art See also Mediagazer
Context & Ripple Effects
The proposal brings a U.S. copyright-disclosure approach closer to an earlier EU AI Act draft that called for training-data copyright disclosure. It also extends a creator-labor debate already visible in the writers’ AI agreement with studios, where training and compensation were negotiated together.
The bill matters because it targets the information gap behind copyright disputes: identifying which protected works entered a model’s training corpus before questions of licensing or compensation can be resolved.
First-order effects
- If enacted, generative-AI developers would have to provide the Copyright Office with a list of copyrighted works used in training and update notices after significant dataset changes.
- Creator and labor groups backing the bill would gain a formal disclosure mechanism; developers, particularly startups, would face new data-tracking and reporting obligations.
Second-order effects
- Training-data provenance would become a more important compliance capability, pushing developers and data suppliers to maintain clearer records of sources, rights status, and dataset revisions.
- Disclosure could give rightsholders more evidence for licensing talks or litigation, while the reported compliance burden may favor better-resourced model developers over smaller entrants.
Third-order effects
- The proposal points toward copyright governance built around training-data transparency rather than leaving model inputs opaque; whether that produces compensation frameworks depends on subsequent legislation, licensing, and court outcomes.
- As creators seek leverage over AI use of their work, collective agreements such as the writers’ studio-AI terms may increasingly be complemented by disclosure rules that apply across developers.
The trend: Generative-AI policy is shifting from broad principles toward auditable rules for the copyrighted material used to build models.