Comments Submitted to Copyright Office on the Right of Publicity and AI
October 26, 2023
In response to the Copyright Office's Notice of Inquiry and Request for Comments on Artificial Intelligence and Copyright, I submitted comments, particularly focused on the Office's questions pertaining to the right of publicity and the use of a person's likeness or performance in the context of generative AI. These comments sweep more broadly than those I submitted last week in response to the circulated discussion draft of a digital replica bill in the U.S. Senate.
In the comments, I observe that there are four primary ways in which people’s identities can be used in generative AI: (1) images, videos, and sound recordings of real people can be used as training data; (2) people’s names and other identifying data can be used as prompts; (3) outputs of generative AI can include, replicate, or evoke the names, likenesses, voices, and performances of identifiable individuals; (4) people’s identities or performances can be used to promote or market AI platforms. My comments, and the concerns highlighted in Congressional hearings on the issue, focus primarily on the third of these uses, in which generative AI produces outputs that are convincing images, sound recordings, or performances that appear to have been done by known and recognizable individuals.
In my comments, I consider how current laws likely extend claims for these outputs, some opportunities for federal legislation in this area, and some potential pitfalls of such legislation that should be avoided. My complete comments can be accessed here: "Artificial Intelligence, Copyright, and Right of Publicity Comments of Professor Jennifer E. Rothman".