Challenges of Artificial Intelligence and Digital Likeness Regulation

The Beginning of the Controversy: Drake's Song and Legislative Absence

New Legal Frameworks: New legal frameworks are moving towards regulating the use of people's faces and voices by Artificial Intelligence. In 2023, the fake track "Heart on My Sleeve," created by AI to mimic artist Drake's voice, sparked widespread controversy. The song was removed from streaming services due to copyright infringement, but the core issue was about likeness impersonation rather than direct content copying.

U.S. Law: U.S. law lacks unified federal legislation on digital likeness issues, with states relying on various laws not originally designed to address advanced AI technologies. However, significant efforts have emerged to change this situation in recent years. In 2024, Tennessee Governor Bill Lee and California Governor Gavin Newsom signed laws to expand protections against unauthorized cloning of entertainment artists, reflecting a legislative shift towards addressing these new challenges.

Technological Evolution and OpenAI

Legal Evolution: However, legal evolution is progressing at a slower pace than technological development. Last October, OpenAI launched its Sora platform, an AI video generation service specifically targeting the capture and remixing of real people's likenesses. This led to the emergence of many astonishingly realistic deepfake videos, including those not approved by their subjects. In response, OpenAI and other companies are implementing their own policies on likeness, which may form new internet rules in the absence of other legislation.

OpenAI Criticisms: OpenAI faced criticism for launching Sora, with CEO Sam Altman claiming it was "too constrained" by safeguards. Nevertheless, the service prompted numerous complaints. After its initial release with few restrictions on the likeness of historical figures, the company reversed its policy following a complaint from Martin Luther King Jr.'s estate regarding "disparaging images" of the civil rights leader. OpenAI also found ways to avoid unauthorized use of living individuals' likenesses, leading to complaints from SAG-AFTRA, prompting OpenAI to strengthen safeguards. Even some individuals who authorized their appearance in Sora (its term for videos using a person's likeness) were disturbed by the results, including "all sorts of salacious outputs" for women.

Legal Challenges and Regulatory Responses

Legal Threats: Potential legal threats concerning unauthorized videos are escalating, with celebrities like Scarlett Johansson seeking legal counsel over the use of their likeness. However, unlike AI copyright infringements, which have led to numerous prominent lawsuits and ongoing deliberations within regulatory bodies, few likeness incidents have escalated to this level, perhaps because the legal landscape remains in flux.

NO FAKES Act: When SAG-AFTRA thanked OpenAI for changing Sora's safeguards, it seized the opportunity to promote the "Nurturing Original Works, Fostering Art, and Keeping Entertainment Safe" (NO FAKES) Act, a years-long effort to codify protections against "unauthorized digital replicas." The NO FAKES Act, supported by YouTube, grants nationwide rights to control the use of a "highly realistic computer-generated electronic representation" of a living or deceased person's voice or visual likeness. The act also includes liability for online services that knowingly permit unauthorized digital replicas.

Criticisms of the NO FAKES Act and the Future of Content

Criticisms of the NO FAKES Act: The NO FAKES Act has drawn strong criticism from online free speech groups. The Electronic Frontier Foundation (EFF) described it as a "new censorship infrastructure" that forces platforms to broadly filter content, inevitably leading to unintended removals and an online "heckler's veto." Social norms are constantly evolving. We are entering a world where you can easily create a video of almost anyone doing almost anything. But the question is: when should you? In many cases, these expectations remain open for discussion.

New Content: More than half of new online content is generated by Artificial Intelligence, raising questions about the nature of authenticity and the challenges of protecting identity from deepfakes.