The American Society of Composers, Authors and Publishers (ASCAP) argued that AI companies need to license material from copyright owners to train their models and that “a new federal right of publicity… is necessary to address the unprecedented scale on which AI tools facilitate the improper use of a creator’s image, likeness, and voice” in a document filed to the Copyright Office on Wednesday (Dec. 6).
The Copyright Office announced that it was studying “the copyright issues raised by generative artificial intelligence” in August and solicited written comments from relevant players in the space. Initial written comments had to be submitted by October 30, while reply comments — which give organizations like ASCAP the chance to push back against assertions made by AI companies like Anthropic and Open AI — were due December 6.
Generative AI models require training: They ingest large amounts of data to identify patterns. “AI training is a computational process of deconstructing existing works for the purpose of modeling mathematically how [they] work,” Google wrote in its reply comments for the Copyright Office. “By taking existing works apart, the algorithm develops a capacity to infer how new ones should be put together” — hence the “generative” part of this.
ASCAP represents songwriters, composers, and music publishers, and its chief concern is that AI companies will be allowed to train models on its members’ works without coming to some sort of licensing arrangement ahead of time. “Numerous comments from AI industry members raise doubts about the technical or economic feasibility of licensing as a model for the authorized use of protected content,” ASCAP writes. “But armchair speculations about the efficiency of licensing do not justify a rampant disregard for creators’ rights.”
ASCAP adds that “numerous large-scale AI tools have already been developed exclusively on the basis of fully licensed or otherwise legally obtained materials” — pointing to Boomy, Stable Audio, Generative AI by Getty Images, and Adobe Firefly — “demonstrating that the development of generative AI technologies need not come at the expense of creators’ rights.”
ASCAP also calls for the implementation of a new federal right-of-publicity law, worried that voice-cloning technology, for example, can threaten artists’ livelihood. “Generative AI technology introduces unprecedented possibilities for the unauthorized use of a creator’s image, likeness, and voice,” ASCAP argues. “The existing patchwork of state laws were not written with this technology in view, and do not adequately protect creators.”
“Without allowing the artists and creators to control their voice and likeness,” ASCAP continues, “this technology will create both consumer confusion and serious financial harm to the original music creators.”