A bill introduced in the US Congress on Tuesday aims to mandate artificial intelligence companies to disclose the copyrighted content they utilize in developing their relational AI models. This legislation responds to the increasing inquiries from lawmakers, journalists, and artists seeking transparency regarding the sources of original works of art—such as songs, physical art, books, and movies—used by AI firms to train their software. There is a growing concern about whether these companies are appropriating copyrighted material without proper authorization to develop their proprietary tools.
The Generative AI Copyright Disclosure Act, proposed by California Democratic senator Adam Schiff, seeks to enforce the submission of copyrighted works by AI firms to the Register of Copyrights before launching new conceptual AI systems that generate text, images, audio, or videos based on user inputs. Failure to comply with this requirement before releasing AI tools to the public could result in financial penalties or render the costs incurred invalid. These submissions may encompass extensive volumes of textual content, graphics, music, or film content.
In a statement, Schiff emphasized the need to strike a balance between the transformative potential of AI and the imperative for social standards and safeguards, stating, “AI has the potential to revolutionize our economy, societal structures, and daily lives.”
There is an ongoing governmental scrutiny into whether major AI corporations, valued in the billions, have infringed upon copyrighted works in violation of international regulations. While Schiff’s bill does not prohibit AI training on copyrighted material, it imposes a significant responsibility on companies to disclose the extensive range of works used in developing tools like ChatGPT, which are typically shrouded in secrecy.
Support for Schiff’s legislation has been voiced by various entities in the entertainment industry, including the Screen Actors Guild of America, Professional Photographers of America, and the American Federation of Television and Radio Artists. Duncan Crabtree-Ireland, SAG-AFTRA’s regional professional director and chief negotiator, emphasized the necessity of protecting intellectual property derived from human creativity, asserting, “Any content generated by AI originates from human creativity.”
Prominent AI enterprises like OpenAI have faced legal challenges alleging the unauthorized use of copyrighted materials in creating tools like ChatGPT. Legal actions have been initiated by individuals such as Sarah Silverman and organizations like the New York Times. OpenAI has bolstered its legal defense team in anticipation of further legal disputes, as reported by the Washington Post.
While denying any wrongdoing, OpenAI and similar AI companies argue that their utilization of copyrighted material aligns with fair use principles, allowing for unregulated use under specific circumstances. This legal stance poses a significant challenge to existing copyright laws, potentially impacting artists’ rights and the financial interests of companies like OpenAI. In a submission to a UK government committee, OpenAI contended that their tools would cease to function effectively without access to copyrighted works.
As generative AI technologies advance, concerns have been raised within the entertainment industry regarding the potential threats to artists’ rights posed by these innovations. Over 200 prominent musicians recently penned an open letter advocating for stricter regulations on artificial intelligence and urging companies to refrain from developing tools that could undermine or replace traditional artistic practices.