AI, Authorship, and the Public Interest Grant Recipients

We’re delighted to announce the five recipients of our AI, Authorship, and the Public Interest grant awards. Chosen from a competitive pool of over 160 proposals, these grantees stood out for their thoughtful, innovative projects that align closely with our research priorities and our mission to serve the public interest.

These grants were made possible by the generous support of the John S. and James L. Knight Foundation. We were impressed by the breadth and quality of the submissions we received. These selected projects reflect a diverse and exciting range of approaches to questions at the intersection of artificial intelligence, authorship, and public good. We look forward to sharing more about their work in the months ahead.

Thank you again to everyone who applied to this grant. We are grateful for your interest and hope to collaborate with many of you in the future.  

Grant Recipients

Public Knowledge: Putting the Public Over Profit: Designing Public Interest Exemptions for Copyright Restrictions on AI Model Development

Public interest AI researchers and developers have distinctly different needs, limitations, motivations, and resources compared to commercial AI companies. Much of the conversation about AI and copyright has flattened the distinction between them. This project aims to begin building a path to protecting public interest AI researchers and developers from copyright restrictions by designing a framework for a public interest exemption against copyright restrictions on AI model development.

1. What is the ideal scope for public interest exemptions (or affirmative protections) from current or future copyright restrictions on AI training?

2. Should public interest exemptions or protections be afforded based on the nature of the use, institutional characteristics, or some other mechanism?
   • How should we address: research organizations, cultural heritage institutions, libraries, open source AI developers, independent researchers, nonprofit developers, and public sector developers?
   • How should downstream commercial uses be handled? How should the US align with the exemptions and limitations created in the EU?
3. How should the US align with the exemptions and limitations created in the EU? Have the EU’s exemptions for research organizations and cultural heritage institutions been effective and sufficiently broad to protect public interest uses and applications of AI training?

Margaret Chon, Seattle University School of Law: Fostering Global Scientific Exchange through GenAI Translations

• This project re-imagines global copyright frameworks to make scientific knowledge available and barrier-free to a majority of the world’s population. Over 90 percent of the current scientific publications are in English but less than 25 percent of the world’s population is proficient in English. Legal barriers to scientific knowledge exchange unnecessarily perpetuate on-going inequities in the distribution of global research. Translation rights are typically assigned by authors to scientific publishers. 

• The project’s intended outcome is widespread action on establishing best legal practices to facilitate accurate and accessible translations of all scientific abstracts (and many full texts) to and from English, thereby expanding global knowledge and technology transfer and facilitating global sustainable development. Its proposed solutions will consist of methods of enhancing legal incentives and reducing legal barriers, with the purpose of expanding translations in specific domains of scientific knowledge, while prioritizing key areas of global public benefits such as human health and climate change research.

Peter Henderson, Princeton University: Between Copyright and Labor

• Generative-AI systems learn from millions of books, songs, and images—sometimes copying enough that authors fear they are being replaced, not just their works being copied. Two recent court rulings (Kadrey v. Meta and Bartz v. Anthropic) highlight the tension: judges say copyright’s “fair-use” test protects public access to new ideas, yet they also worry about eroding the livelihoods of human creators. This project asks an underlying question: Does U.S. copyright law shield creative workers (or humans as a whole) from AI-driven competition, or does it only guard the market value of their individual works, or none of these? If not, what alternative pathways exist to incentivize creative labor?

• This project will answer in three steps. First, it will dig through a century of legislation and court opinions to see how the law has treated copyright market impacts, particularly as related to protecting human creative labor. Second, it will examine the newest AI cases to identify what evidence plaintiffs must provide when they claim AI systems undercut human jobs. Third, it will build an economic model that projects how different policy options—status-quo fair use, a licensing levy on AI training, or new tax-and-spend programs—might affect wages and incentives for human creative output.

Jacob Noti-Victor, Cardozo Law School: Dissecting the AI-Assisted Work: Copyright Infringement Litigation in the Age of Generative AI

• This project examines the copyrightability of AI-generated materials from the perspective of copyright infringement doctrines and policies. The Copyright Office has generally declined to protect AI-generated materials but now allows some limited protection when human selection, coordination, or arrangement is involved. An undiscussed byproduct of this approach is that it will significantly complicate substantial similarity analysis in future cases involving AI-assisted works and infringing copies. The highly expressive yet unprotectable AI-generated elements must be meticulously distinguished from human-authored contributions, and judges will only be able to achieve this task by examining the likely voluminous record of a work’s creation. This complexity not only threatens to burden judicial administration but also challenges established substantial similarity doctrine.

• The project explores doctrinal solutions that aim to streamline judicial decision making, possibly borrowing from special tests applied in contexts such as software and design patents. It further investigates technical measures, including the use of AI tools, to simplify distinguishing between AI-generated and human-contributed expressions. Finally, it investigates the broader copyright policy questions implicated by these challenges. 

Association of Moving Image Archivists: Preserving Trust: Archives, Generative AI, and the Public Interest

• This project investigates how generative AI is reshaping the work of audiovisual archives, vital to preserving truth, cultural memory, and public trust. As AI-generated and AI-altered media become more sophisticated and widespread, archives face mounting legal, ethical, and technical challenges. To uphold their role as trusted stewards of history, archives must develop new strategies for appraisal, documentation, and authentication while actively defending against manipulation, misinformation, and the erosion of public trust.

• Led by the Association of Moving Image Archivists (AMIA) and the Archival Producers Alliance (APA), this project brings together archivists, technologists, copyright experts, and media makers. Through collaborative working groups, fieldwide surveys, and public roundtables, the project will examine legal, ethical, and practical questions surrounding provenance, licensing, fair use, and institutional readiness.