In a submission to the Copyright Office on Wednesday, the American Society of Composers, Authors, and Publishers (ASCAP) contended that AI companies must secure licenses from copyright holders to train their models. They also advocated for the necessity of a new federal right of publicity to tackle the extensive misuse of images, likenesses, and voices enabled by AI tools.
The Copyright Office initiated research on “the copyright issues raised by conceptual artificial intelligence” in August, soliciting feedback from relevant industry stakeholders. Initial written opinions were due by October 30, with email responses, allowing entities like ASCAP to counter assertions from AI firms such as Anthropic and Open AI, required by December 6.
ASCAP’s primary apprehension, representing musicians, composers, and music publishers, revolves around the potential unauthorized utilization of their members’ creations by AI entities for model training without prior negotiation of licensing terms. They highlight that while some AI industry comments question the feasibility of licensing in terms of technical or financial aspects, mere speculative discussions do not justify disregarding authors’ rights.
Google, in its feedback to the Copyright Office, elucidated that AI education involves deconstructing existing functions mathematically to comprehend their operational principles. The “generative” aspect entails the algorithm acquiring the ability to infer the construction of new works by analyzing existing ones.
ASCAP underscores the importance of ensuring that AI systems are developed using fully licensed or legitimately obtained materials, citing examples like Boomy, Stable Audio, Generative AI by Getty Images, and Adobe Firefly. These instances illustrate that the advancement of relational AI systems can coexist with authors’ rights without compromise.
Furthermore, ASCAP advocates for the enactment of a new national right-of-publicity law to address concerns regarding the potential threats posed to artists’ livelihoods by technologies like voice-cloning systems. They emphasize that the current patchwork of state laws is insufficient to safeguard creators against the unauthorized exploitation of their images, likenesses, and voices facilitated by Generative AI systems.
ASCAP warns that the proliferation of this technology could lead to consumer confusion and substantial economic losses for original music creators, emphasizing the necessity for artists and creators to retain control over their voices and likenesses.
