ASCAP Doubles Down on Support for ‘Direct Voluntary Licensing,’ Right of Publicity Expansion in New Copyright Office AI Comments

ASCAP AI
  • Save

ASCAP AI
  • Save
Photo Credit: ASCAP

The American Society of Composers, Authors and Publishers (ASCAP) has submitted new comments on artificial intelligence to the Copyright Office, addressing the arguments of AI and tech companies including Anthropic.

Update: Our original title (published Friday, December 9th) used the licensing term ‘direct licensing,’ using language from the submission, though an ASCAP representative requested that we change this to ‘voluntary licensing’ given that ‘direct licensing’ is ‘inaccurate.’ We have since updated this to ‘direct voluntary licensing’ to reflect the exact language of the document.

ASCAP just recently filed this more concise reply, having laid out in detail its position on and concerns about the unprecedented technology roughly one month ago. These remarks (like those from the aforementioned Anthropic, AI music generator Boomy, and several others) were delivered in connection with the Copyright Office’s August 30th request for comment on AI and copyright.

Somewhat surprisingly, ASCAP’s latest submission doesn’t focus on the prevalent claim that training AI systems on protected materials constitutes fair use. Rather, the text centers on three main points, the first being the broader importance of direct licensing for creators. (An ASCAP representative later told us that the organization had contested the fair use arguments in an earlier filing).

“[A]rmchair speculations about the efficiency of licensing do not justify a rampant disregard for creators’ rights,” the PRO relayed of certain parties’ opposition to licensing. “As they have done countless times in the past, licensing models will adapt to the evolving technical environment to ensure that creators are compensated for the use of their intellectual property.”

Behind the stance, ASCAP also took aim at the idea “that the sheer volume of training data required for developing AI tools generally precludes direct voluntary licensing.” And in support of the point, the non-profit cited Boomy, or an AI platform “developed exclusively on the basis of fully licensed or otherwise legally obtained materials.”

Next, ASCAP emphasized that AI, contrasting “innovations like sound mixing, autotune,” and more, “poses the very real threat of supplanting—rather than supporting—human creativity.”

“No previous technology can do what generative AI has made possible: generating new content; near autonomously; at large scale; instantaneously; and at a level of quality that is increasingly indistinguishable from human work,” the organization spelled out. “The resulting threat to creators’ livelihoods is not idle speculation, but a real and growing phenomenon.”

Consequently, it’s imperative that the potential advantages of AI be balanced “against their very real exploitative and displacing effects on the human creators without which they could not exist,” according to the text.

Lastly, ASCAP stressed the perceived significance of an enhanced federal right of publicity, maintaining that arguments against the bolstered protection overlook AI’s “unprecedented possibilities for the unauthorized use of a creator’s image, likeness, and voice” as well as the purported inadequacy of “the existing patchwork of state laws.”

One such bill, the No Fakes Act, debuted in October and, per its discussion draft, would establish penalties for any unapproved “digital replica involving a sound recording artist.” As described by the involved lawmakers, “digital replica” refers to any computer-generated “image, voice, or visual likeness of an individual” that’s “nearly indistinguishable” from the original.

Of course, it’ll be worth following the progress of the legislation as it takes shape and potentially picks up steam in the approaching months. Among different considerations, it’s unclear precisely how the bill would (or whether it could) address digital replicas of original voices that naturally resemble other, commercially prominent original voices – especially given the availability of these replicas’ training data.