• Federal courts, not the U.S. Copyright Office, will determine how AI training on copyrighted works aligns with fair use.

• The Copyright Office’s report is expected to provide a framework that could guide courts, policymakers, and stakeholders.

• The debate over AI and copyright is likely to reach the Supreme Court, with Congress potentially stepping in to legislate.

The future of artificial intelligence and fair use copyright hinges on federal judges, not the U.S. Copyright Office, a Bloomberg Law report notes.

While the office is compiling public feedback and releasing an analytical report, it doesn’t have the authority to interpret copyright infringement law. That responsibility lies with the courts, which are currently handling dozens of lawsuits involving AI companies like OpenAI, Microsoft, and Meta.

Judges will use the Copyright Act’s four-factor test for fair use to decide whether training AI models on copyrighted material without permission is lawful. These factors include:

• the purpose and transformativeness of the use,

• the nature of the original work,

• the portion of the work used,

• the potential market impact.

The Copyright Office’s Role

The office released the first part of its fair use copyright report in July, calling on Congress to pass a law protecting people from unauthorized deepfakes. 

Register of Copyrights Shira Perlmutter recently told senators the office plans to publish by the end of 2024 the two remaining sections of the report and to provide a framework for analyzing the fair use doctrine’s applicability to AI’s use of copyrighted works.

Perlmutter told lawmakers that while the office’s findings won’t dictate court decisions, they could provide valuable insight.

• Santa Clara University law professor Edward Lee expects the office to avoid taking categorical stances. Instead, its framework might act like an amicus brief for courts, synthesizing stakeholder perspectives and technical details about how generative AI works.

• Greenberg Glusker partner Aaron Moss noted that the report could serve as a roadmap for courts grappling with these complex issues. “This is the first time we’re collecting information from all the various stakeholders,” Moss said, emphasizing the potential value of the report’s synthesis of views.

Public Comments and Legal Disputes

The Copyright Office has received over 10,000 public comments on AI and copyright issues. Opinions diverge sharply:

• AI Companies argue that training models on copyrighted works constitutes transformative fair use. OpenAI likens this to creating something new from existing material.

• Creators and Rights Holders contend that AI’s use of copyrighted material often leads to illegal reproductions, undermining original works. The Authors Guild has called for legislative intervention if courts fail to preserve copyright incentives.

• Legislators like Sen. Marsha Blackburn believe unlicensed use should not qualify as fair use when AI-generated outputs compete commercially with human-made content.

The tension between innovation and the protection of intellectual property has spurred litigation, with OpenAI, Meta, and others defending their practices in ongoing lawsuits.

Will Congress Step In?

While courts will have the first say, the report’s findings could influence future legislation. Perlmutter acknowledged that court rulings might drive stakeholders to Congress for clarification or change.

Michael Graif of Brown Rudnick expects lawmakers to rely heavily on the Copyright Office’s expertise if legislative action becomes necessary.

OpenAI has urged caution, warning that premature legislation could stifle technological advancement. However, the Authors Guild and others argue that gaps left by the courts could harm creative industries, necessitating Congressional intervention.