WASHINGTON — Senators Adam Schiff, a Democrat from California, and John Curtis, a Republican from Utah, introduced bipartisan legislation on Thursday aimed at increasing transparency in the artificial intelligence industry by requiring companies to disclose the copyrighted materials used to train their AI models. The bill, known as the Copyright Labeling and Ethical AI Reporting Act, or CLEAR Act, targets a contentious issue at the intersection of Hollywood and Silicon Valley, where creators have raised alarms about the unauthorized use of their works in developing generative AI technologies.
The legislation would mandate that AI developers file a detailed notice with the Register of Copyrights before releasing a new model to the public. This notice must outline the copyrighted works incorporated into the training datasets. Notably, the requirement applies retroactively to models already available to consumers, meaning companies like OpenAI, Google and others with existing products would need to comply by submitting the information for their current systems.
Under the bill, the U.S. Copyright Office would be tasked with creating and maintaining a public database of these disclosures, allowing creators, researchers and the public to access information on how AI models are built. Failure to file the required notice could result in civil penalties, though the exact amounts and enforcement mechanisms would be determined by the Copyright Office.
The introduction of the CLEAR Act comes amid growing tensions in the entertainment industry over AI's role in content creation. For years, writers, actors, musicians and visual artists have expressed concerns that large language models and image generators are being trained on vast troves of copyrighted material scraped from the internet without permission or compensation. This practice, proponents of the bill argue, undermines the livelihoods of creative professionals who form the backbone of America's cultural economy.
Several prominent unions and guilds have thrown their support behind the measure. Among the endorsers are SAG-AFTRA, which represents actors and other media professionals; the Writers Guild of America, both West and East branches; the Directors Guild of America; the International Alliance of Theatrical Stage Employees, or IATSE; the American Federation of Musicians; the Authors Guild; the National Association of Voice Actors; the National Music Publishers' Association; the Artists Rights Alliance; SoundExchange; and the Television Academy. These groups, which collectively represent hundreds of thousands of workers in the creative sectors, see the bill as a crucial step toward protecting intellectual property in the AI era.
The Recording Industry Association of America, or RIAA, has also voiced its backing for the legislation, highlighting the music industry's particular vulnerability to AI-generated content that mimics artists' styles. However, the Motion Picture Association, which represents major studios like Warner Bros., Disney and Universal, is conspicuously absent from the list of supporters. Industry insiders point to internal divisions among studios on how to approach AI, with some executives believing that existing copyright laws are sufficient to address infringement claims without new mandates.
"While AI has the potential to improve our lives and change the way we work and innovate, we need a unified approach to implementing guardrails that protect the work and livelihoods of all workers, including artists and creators," Schiff said in a statement. "Human creativity is the foundation of our cultural and creative economy, and it plays a vital role in shaping our society, our stories, and our shared experiences."
Curtis echoed this sentiment, emphasizing the need for balance in fostering innovation. "Congress must help encourage AI innovation, but not without transparency and accountability," he said. "The CLEAR Act strikes the right balance by protecting creators’ intellectual property while providing clear expectations for companies. By shedding light on how generative AI models are trained, our bipartisan legislation will help build public trust for emerging technologies and foster the best of American creativity."
The bill does not go as far as requiring AI companies to obtain licenses for the copyrighted works used in training, a more aggressive measure that has been the focus of ongoing litigation. Multiple lawsuits have been filed by authors, news outlets and individual creators alleging that AI firms like Stability AI and Midjourney have infringed copyrights by training on protected materials without permission. In one high-profile case, Warner Bros., The Walt Disney Co. and NBCUniversal have sued Midjourney, claiming the AI image generator used their copyrighted characters and visuals in its development process.
AI companies, for their part, have defended their practices as fair use under current copyright law, arguing that training models on public data is transformative and essential for technological advancement. This defense has been central to cases like the New York Times' suit against OpenAI and Microsoft, where the newspaper accused the companies of using millions of its articles to build ChatGPT without authorization.
Former President Donald Trump, who has commented on AI policy during his current campaign, has weighed in on the licensing debate, calling such requirements impractical. In remarks earlier this year, Trump suggested that mandating licenses for training data would stifle innovation and burden startups, though he has not directly addressed the CLEAR Act.
The entertainment industry's unease with AI predates this bill and has influenced recent labor negotiations. Just last year, SAG-AFTRA's strike centered on AI protections, with actors demanding consent and compensation for the use of their likenesses in digital replicas. Similarly, the Writers Guild secured provisions in their contract to regulate AI's role in scriptwriting. These battles underscore the broader context in which the CLEAR Act emerges, as Hollywood grapples with a technology that could automate jobs while generating new revenue streams.
Proponents of the bill argue that transparency alone could pave the way for fairer practices. By making training data public, creators could more easily identify infringements and pursue remedies under existing law, potentially reducing the need for widespread licensing mandates. Critics within the tech sector, however, worry that disclosure requirements could expose trade secrets or invite frivolous lawsuits, slowing the pace of AI development at a time when the U.S. seeks to compete with global rivals like China.
As the bill moves through Congress, its fate remains uncertain in a divided Senate. Schiff and Curtis, representing opposite parties, hope the bipartisan nature of the proposal will garner broader support. The legislation has already sparked discussions in committee hearings, where witnesses from both the creative and tech communities are expected to testify in the coming weeks.
Looking ahead, the CLEAR Act could set a precedent for regulating AI beyond copyright issues, touching on ethics, bias and national security. For the creative industries, passage of the bill would signal a congressional acknowledgment of their concerns, potentially leading to more comprehensive reforms. Meanwhile, AI enthusiasts caution that overregulation might hinder the transformative potential of these tools, which have applications in medicine, education and beyond.
In the end, the debate encapsulated by this legislation reflects a pivotal moment for American innovation: balancing the protection of human ingenuity with the promise of machine learning. As lawmakers deliberate, the eyes of Hollywood and Silicon Valley alike will be on Capitol Hill, watching to see if transparency can bridge the divide.