Singers, authors specializing in romantic fiction, video game developers, and voice actors are urgently appealing to the U.S. government for assistance in mitigating the threat posed by artificial intelligence to their professions.
Expressing apprehension, a blogger emphasized, “Choose control AI.” He voiced concerns that AI might replicate his content, reflecting the sentiments echoed in the multitude of letters recently submitted to the U.S. Copyright Office.
In contrast, technology enterprises seem largely satisfied with the current situation as it enables them to consume published content to enhance the capabilities of their AI systems in emulating human behavior.
The highest-ranking rights official in the nation has not yet taken a definitive stance. While contemplating the necessity for rights reforms in response to the emergence of advanced AI tools generating compelling multimedia content, she assured The Associated Press that she is attentive to all viewpoints.
During an interview, Shira Perlmutter, the head of the U.S. Copyright Office, disclosed, “We have received nearly 10,000 comments,” emphasizing that each submission is meticulously reviewed by human evaluators rather than automated processes. She personally dedicates significant effort to scrutinizing a substantial portion of the feedback.
WHAT’S AT STAKE?
Perlmutter oversees the U.S. Copyright Office, which recorded over 480,000 rights registrations last year, encompassing numerous creative works. However, the office is increasingly confronted with requests to register AI-generated content. Given that copyright laws are designed to protect human-authored works, claims involving entirely machine-generated content have generally been dismissed.
An essential query posed by Perlmutter is whether human involvement in directing the output of AI systems reaches a threshold where the individual can be deemed a contributor to the authorship, as humans input data into AI models and provide guidelines to influence the outcomes.
One pertinent issue raised by the Copyright Office is the treatment of copyrighted works sourced from the internet and other outlets for training AI models, often without authorization or compensation. This matter has elicited extensive feedback from creative professionals.
Following the closure of the initial comment period in early October, over 9,700 remarks were submitted to the Copyright Office, a division of the Library of Congress. A subsequent round of responses is scheduled by December 6, after which Perlmutter’s team will formulate recommendations for Congress and other stakeholders regarding potential reforms.
ARTISTIC COMMUNITY SENTIMENTS
Renowned actor and director Justine Bateman expressed dismay over AI systems “ingesting 100 years of film” and television content, potentially disrupting the cinematic landscape and altering significant aspects of the entertainment industry. Bateman urged the U.S. Copyright Office to intervene and halt what she views as a substantial violation of creative rights.
Television showrunner Lilla Zuckerman advocated for industry action against AI technologies, characterizing them as “nothing more than a plagiarism machine” that could jeopardize the integrity of artistic endeavors. Similarly, state singer Marc Beeson highlighted the risks facing the music sector, cautioning against the unregulated use of AI that could undermine a quintessential form of American art.
While individual commentators expressed shared concerns, major music labels like Universal Music Group and prominent media entities such as the New York Times and The Associated Press also echoed apprehensions regarding AI’s impact on creative industries.
LEGAL PERSPECTIVES
Major tech corporations including Google, Microsoft, and OpenAI, the developers of ChatGPT, assert to the Copyright Office that their AI training practices align with the “fair use” doctrine, permitting limited utilization of copyrighted materials for purposes such as education, research, or transformative creation.
Meta Platforms, the parent company of social media giants like Facebook and Instagram, underscored that AI training in the UK leverages the permissibility of using copyrighted content for educating generative AI models. They clarified that the objective is to identify patterns across diverse content sets, not to replicate or reproduce specific works.
Court rulings thus far have predominantly favored technology firms in navigating copyright laws concerning AI systems. While a federal judge in San Francisco recently dismissed a substantial portion of a legal challenge against AI image generation, certain claims were allowed to proceed, signaling a partial setback for traditional artists.
The Copyright Office’s ongoing efforts aim to address these complex issues, distinguishing them from past legal precedents such as the Google Books case. Heidi Bond, a former law professor and bestselling romance author known as Courtney Milan, contended that while fair use encompasses extracting knowledge from books, Google Books obtained lawful copies from libraries and institutions, unlike some AI developers engaging in unauthorized replication of literary works.
Perlmutter emphasized the distinctive nature of the current challenges compared to past cases, highlighting the nuanced considerations surrounding fair use protections in the realm of AI-generated content.
