The House Judiciary Subcommittee on Courts, Intellectual Property, and the Internet held a hearing on Wednesday, May 17, 2023, at 10:00 a.m. ET. The hearing, "Artificial Intelligence and Intellectual Property: Part I — Interoperability of AI and Copyright Law," will examine the intersection of generative artificial intelligence (AI) technology and copyright law. Among the topics that will be addressed are the use of copyright-protected works in the training of generative AI models, copyright protection of works that were produced with the assistance of generative AI, and the economic impact of generative AI on creators and creative industries.
Jeff Sedlik's written testimony:
Chairman Issa, Ranking Member Johnson, and other Members of the Subcommittee on Courts, Intellectual Property, and the Internet, thank you for the opportunity to testify today on Artificial Intelligence and Intellectual Property.
My name is Jeffrey Sedlik. I have been a professional advertising photographer and filmmaker for 37 years, and am the president of the PLUS Coalition at www.PLUS.org, a global non-profit organization in which diverse stakeholders from 140 countries are collaborating to develop a global registry for the identification of visual works. I previously served as president of the American Photographic Artists.
I am the creator and owner of hundreds of thousands of copyrighted photographs. I make my living by creating and licensing my photographs to appear in all manner of media, for all manner of purposes, by all manner of clients. Like all visual artists, my ability to create new works, sustain my business, and support my family depends directly on my exclusive rights to reproduce, distribute, publicly display, and adapt my original works. These fundamental rights, guaranteed to photographers since President Lincoln authorized the extension of copyright protection to photographs in 1865, are the core of my business, providing a strong incentive to create new works, just as envisioned by the Framers in Article 1, Section 8, Clause 8 of the Constitution.
We photographers have never shied away from technology. In fact, we have a 197-year history of embracing technology as a vehicle for visual storytelling. We tell our stories with original, human expression.
Professional photographers don’t “snap” photographs, nor are we mere button pushers. In preparing to create an image, a professional photographer determines the optimal vantage point, carefully selecting foreground, background, natural lighting, artiﬁcial lighting, shapes, textures, perspectives, storytelling potential, and numerous other factors. Our creative decisions are guided by our life history, and our unique combination of training, experience, personality, aesthetic sensibility memories, research, and other factors. We anticipate and respond dynamically to the subject matter and shooting environment; we exercise control over the visual rendition of the scene; we decide which elements to include or exclude, and where to place those elements within the frame; we determine how to juxtapose people, objects, and other compositional elements for desired effect; we control the placement and interplay of color, tone, contrast, light, and shade; we control perspective, distortion, depth of ﬁeld, and selective focus to guide the viewer’s eye through the photograph; and we select the precise moment at which to create the ultimate photograph.
For some photographers, these creative decisions may occur over a period of days, weeks, or months. For others, such as photojournalists, thousands of creative decisions are by necessity compressed into minutes, even seconds.Irrespective of the time dedicated, photographers in all genres exercise original human expression, subjectively selecting, coordinating and arranging the elements of our creative works.
Copyright law affords protection to original human expression. While creators can and do employ AI technologies as a starting point for original human expression, and while that human expression should be eligible for copyright protection, the output of an image by a generative AI platform in response to prompts, in the manner employed today, is not original human expression, and should remain ineligible for copyright protection.
Most AI developers and platforms build their businesses by exploiting billions of creators’ works without authorization from the copyright owners. Many AI developers claim that ingestion of copyrighted works is fair use. Fair use is an affirmative defense, not a right. A determination of fair use is a fact specific inquiry that must be made on a case-by-case basis. The reproduction right under Section 106 of the Copyright Act is a standalone right. The creation and use of copies of protected works for AI ingestion and generation purposes is copyright infringement, on a massive scale.
AI developers must be required not only to obtain advance permission to ingest and exploit creative works, but to also compensate creators. Policies must not favor machines over human creators. All rights of human creators and copyright owners must be respected. Available compensation schemes include direct compensation to artists, compensation through licensed agents, and collective licensing schemes managed by copyright management organizations, such as the American Society for Collective Rights Licensing.
In the photography industry, we have a long-standing tradition of offering and selling licenses for “artist reference.” These licenses typically permit the use of a photograph by an artist as a visual reference for the creation of one or more derivative works. The ingestion of a photograph by an AI system for the purpose of reference in the output of machine generated images is a form of reference of the expression in the photograph. In fact, stock photography agencies now routinely sell licenses to AI platforms for the purpose of AI ingestion and generative AI use.
Photographers increasingly offer similar licenses. Clearly, a market exists for these licenses. The unlicensed ingestion of photographs by AI systems usurps that market.
The high-quality works created by photographers and other visual artists are ideal for ingestion by AI machines to generate output. AI developers use these works because of their signiﬁcant expressive value. Most of these works are scraped from websites without the knowledge or permission of the copyright owners. I have identiﬁed thousands of unlicensed copies of my works in the databases used by AI systems to support image generation. My peers report similar ﬁndings. It is clear that many AI platforms are trained on infringing copies of creative works, and that generative AI is a technology founded on copyright infringement.
It is essential that AI technologies be developed and used in a manner that is responsible, respectful, and ethical – and in a way that upholds the underlying goals and purposes of our copyright system.
Thank you for your efforts to ensure that creative works are protected, and for the opportunity to share my experience and perspective in my testimony today.
Jeffrey Sedlik, Sedlik Photography, PLUS Coalition, American Photographic Artists