Developments in the artificial intelligence (“AI”) space dominated (again) this week. First up is the latest rejection from the U.S. Copyright Office (“USCO”), which has refused to register an AI-generated artwork. In a Sept. 5 letter, the Copyright Review Board responded to Colorado-based artist Jason Allen’s Second Request for Reconsideration for Refusal to Register “Théâtre D’opéra Spatial,” a 2D artwork that Allen created using Midjourney.
Some Background: “While Mr. Allen did not disclose in his application that the work was created using an AI system, the Office was aware of the work because it had garnered national attention for being the first AI-generated image to win the 2022 Colorado State Fair’s annual fine art competition,” the Board stated in its letter. In response to a request for additional information about Allen’s use of text-to-image generator, Midjourney, the Board states that “Mr. Allen provided an explanation of his process, stating that he “input numerous revisions and text prompts at least 624 times to arrive at the initial version of the image. He further explained that, after Midjourney produced the initial version of the Work, he used Adobe Photoshop to remove flaws and create new visual content and used Gigapixel AI to ‘upscale’ the image, increasing its resolution and size.” As a result of these disclosures, the USCO’s the examiner requested that the features of the work generated by Midjourney be excluded from the copyright claim.
When Allen refused the examiner’s request and reasserted his claim to copyright in the features of the work produced by an AI system, the USCO refused to register the claim because the deposit for the work did not “fix only [Allen’s] alleged authorship” but instead included ‘inextricably merged, inseparable contributions’ from both Mr. Allen and Midjourney.”
In Jan. 2023, Allen requested that the USCO reconsider its initial refusal to register the work, arguing that the examiner had misapplied the human authorship requirement and that public policy favored registration.
Following an initial review, the USCO concluded that the work could not be registered without Allen limiting the claim to only the copyrightable authorship he contributed to the work. The crux of the USCO’s pushback: “The image generated by Midjourney that formed the initial basis for the [Théâtre D’opéra Spatia] is not an original work of authorship protected by copyright.” The USCO accepted Mr. Allen’s claim that human-authored “visual edits” made with Adobe Photoshop contained a sufficient amount of original authorship to be registered. However, it explained that the features generated by Midjourney and AI-powered image upscaler Gigapixel AI must be excluded as non-human authorship. Because Mr. Allen sought to register the entire work and refused to disclaim the portions attributable to AI, the USCO refused to register the claim.
In a second bid for reconsideration, Allen put forth three key arguments …
(1) “Regardless of whether the underlying AI-generated work is eligible for copyright registration, the entire work in the form submitted to the copyright office should be accepted for registration” because the underlying AI-generated work “merely constitutes raw material which Allen has transformed through his artistic contributions.”
(2) By refusing to register content generated via Midjourney and other generative AI platforms, “the Office is placing a value judgment on the utility of various tools,” and that denial of copyright protection for the output of such tools would result in a void of ownership.
(3) Allen objected to the USCO’s registration requirements for works containing AI-generated content, stating that “[r]equiring creators to list each tool and the proportion of the work created with the tool would have a burdensome effect if enforced uniformly.”
The USCO’s Second Refusal: After examining the work and considering the arguments made in Allen’s First and Second Requests, the Board upheld the USCO’s initial determination, finding that the work “contains more than a de minimis amount of AI-generated content, which must be disclaimed in an application for registration.” Because Allen has refused to disclaim the material produced by AI, the work cannot be registered as submitted, the Board confirmed.
In the wake of the rejection from the Board, Allen said he is “certain we will win in the end,” noting that if the Board’s decision stands, it will “create more problems than it solves. This is going to create new and creative problems for the copyright office in ways we can’t even speculate yet.”
> A Recent Entrance to Paradise: The USCO’s determination follows closely from a decision from a district court in DC, which held that AI-created artwork is not eligible for copyright protection, thereby, confirming a decision from the USCO to deny Stephen Thaler a copyright registration for a 2D work, “A Recent Entrance to Paradise,” created by an AI program.
> Zarya the Dawn: The USCO determined in February that “Zarya of the Dawn,” a graphic novel “comprised of human-authored text combined with images generated by the AI service Midjourney,” constitutes “a copyrightable work, but that the individual [AI-generated] images themselves cannot be protected by copyright.”
> USCO Policy Statement: In March, Register of Copyrights and Director of the U.S. Copyright Office, Shira Perlmutter, said in a statement that in the Office’s view, “it is well-established that copyright can protect only material that is the product of human creativity,” and that “most fundamentally, the term ‘author,’ which is used in both the Constitution and the Copyright Act, excludes non-humans.”
> USCO Call for Comments: In late August, the USCO announced that it is seeking comments in connection with a study of the law and policy issues posed by AI systems amid the rise of complicated questions about authorship and infringement, and as a growing number of copyright-centric lawsuits are being waged against generative AI giants like OpenAI. According to the notice published in the Federal Register on Aug. 30, the USCO states that it is seeking insight on “the use of copyrighted works to train AI models, the appropriate levels of transparency and disclosure with respect to the use of copyrighted works, and the legal status of AI-generated outputs” not only to inform its study and help it assess whether legislative or regulatory steps in this area are warranted.
Senators Richard Blumenthal (D-CT) and Josh Hawley (R-Mo.) are preparing to announce a sweeping framework to regulate artificial intelligence. The New York Times reported this week that the two leaders of the Senate judiciary’s subcommittee for privacy, technology and law “said in interviews on Thursday that their framework will include requirements for the licensing and auditing of A.I., the creation of an independent federal office to oversee the technology, liability for companies for privacy and civil rights violations, and requirements for data transparency and safety standards.”
Until further details about the senators’ AI-focused framework are available, you can find our AI Legislation Tracker right here.
“Some customers are concerned about the risk of IP infringement claims if they use the output produced by generative AI, [which] is understandable, given recent public inquiries by authors and artists regarding how their own work is being used in conjunction with AI models and services,” Microsoft said in a blog spot this week. “To address this customer concern, Microsoft is announcing our new Copilot Copyright Commitment,” which will see the tech titan “assume responsibility for the potential legal risks involved” in connection with parties’ use of the AI assistant feature for Microsoft 365 applications and services.
Not the first to make such a pledge, Microsoft’s “Copilot Copyright Commitment” follows from Adobe’s June 2023 announcement that its generative AI platform Adobe Firefly – which is describes as “Generative AI for creatives” – is “designed to be commercially safe and trained on licensed Adobe Stock and public domain images where copyright has expired, so your enterprise can generate content for public and commercial use with higher peace of mind.” In particular, Abode says that “enterprises also have the opportunity to obtain an IP indemnity from Adobe for content generated by select workflows powered by Firefly.” According to ComputerWorld, “The basic idea of the Stock indemnification is that, as long as the customer is using the product within the terms and conditions, Adobe will compensate the customer for any IP-related legal claims that might arise.” An spokesperson stated at the time of the initial announcement that the details for Firefly were still finalized, with a full launch of Firefly for Enterprise offering not expected until the second half of this year.
A month later, Shutterstock revealed that that it will provide Enterprise customers with “full indemnification for the license and use of generative AI images on its platform.” The company, which launched its AI image “Generate” tool in January, says that it “will fulfill indemnification requests on demand via human review with the intent to protect its customers against potential claims related to their use of generative AI images created and licensed on shutterstock.com.