The U.S. Constitution provides the basis for copyright protection “by securing for limited times to authors . . . the exclusive right to their respective writings” in order “[t]o promote the progress of . . . the useful arts.” But when it comes to images created by artificial intelligence (AI), who is the author? And is the AI-generated image copyrightable at all?
In his ongoing efforts to secure intellectual property rights for the creations of his AI machine, DABUS—which we’ve previously written about here and here—Dr. Stephen Thaler is now asking a federal court to find that not only is DABUS’s work copyrightable, but that he is the copyright owner by virtue of having created DABUS and having told it what to do.
In an at-times academically dense summary judgment motion filed last week, Dr. Thaler advocates for the copyrightability and registrablity of an AI-generated image titled A Recent Entrance to Paradise. The image, which Thaler has described as a “simulated near-death experience,” was generated in 2012 by Thaler’s “Creativity Machine” (the moniker given DABUS in the motion). Thaler applied for a copyright registration in 2016 but was denied by the Copyright Office twice. According to Thaler’s motion, in doing so the Copyright Office relied on cases “from the 19th century [that] greatly predate even the invention of the first modern computers” to impermissibly impose a “Human Authorship Requirement.” Since the Creativity Machine is not a human, Paradise is not copyrightable according to the Copyright Office.
Not so, says Thaler’s motion. According to Thaler, Paradise and other AI-generated images are copyrightable because the AI qualifies as an “author” under the Copyright Act under the plain and ordinary meaning of “author,” which, he argues, need not be human. Noting that corporations can be “authors” under the Act’s work made for hire doctrine, Thaler argued “author” does not necessarily mean “human.” And, he argued, finding an AI machine to be an “author” under the Copyright Act promotes the arts not by motivating the AI to generate additional works, but rather by motivating “individuals like Dr. Thaler, and businesses like music and movie studios, . . . to develop and use AI to generate new works, thereby achieving the purpose of the [A]ct.”
Turning to the issue of ownership, Thaler rejected the Copyright Office’s conclusion that even if Thaler’s Creativity Machine could generate a copyrightable work it could not transfer the copyright to Dr. Thaler as “fail[ing] to recognize two separate mechanisms that would make Dr. Thaler the original owner by operation of law given standard property principles.” First, he argued, because he owns the Creativity Machine, by operation of law he owns the copyright in whatever it generates. Analogizing to a fruit tree, Thaler argued that if he owned that tree, he would also own its fruit; “[t]his does not require the tree to execute a written agreement to transfer the fruit, the fruit belongs to Dr. Thaler by virtue of his relationship to the tree.” Alternatively, under an approach that could pejoratively be referred to as “finders keepers,” Thaler argued he owns the copyright in anything generated by his AI “by virtue of being the first party to possess it and communicate his ownership.”
Perhaps sensing these two approaches derived from traditional property law might be a bit tough for a court facing an intellectual property case to swallow, Thaler also argued authorship and ownership in Paradise could be found to vest in him automatically under the Copyright Act’s work made for hire doctrine. According to Thaler, the doctrine applies because the Creativity Machine created Paradise at Thaler’s direction and instruction, created it for Thaler, and could be considered an “employee” of Thaler under the plain and ordinary meaning of employee.
While interesting, Thaler’s motion seems to lead to more questions than it does answers. Without a doubt, traditionally human-created works are being and will continue to be created by AI. And there’s some persuasive force behind the idea that AI-created images could be copyrightable. But AI cannot be the owner of the copyright, at least not in any meaningful sense. Even Thaler’s motion recognizes that if AI were the copyright owner it would lack standing to sue based on Ninth Circuit decisions holding a monkey lacked standing to bring a copyright infringement claim. If the owner is instead the person who created the AI, who owns images created by the AI at the direction of someone other than the AI’s creator? Social media has been flooded the last few months with examples of AI-generated user portraits; clearly those portraits were not all created at the direction of the AI’s creator but rather by users of the creator’s AI. If the “owner” is instead the person that directs the AI to create an image, like a social media user, does that mean that there’s no “incentive” to machine creators, music and movie studios, and the like to create AI in the first place?
These and other questions are unlikely to be resolved anytime soon. Thaler’s efforts in the AI intellectual property space have been described as “an academic project,” reportedly underway on a pro bono basis by lawyers that see these as important legal issues in need of resolution. Thaler has only had limited success to date abroad, and no success in the United States. But as Thaler’s motion argues, courts interpret the Copyright Act in light of “significant changes in technology.” As his motion points out, “[n]ow that, as a factual matter, the universe of authorship has expanded to include AI, there is a need to consider how the purpose of the Act can best be achieved in light of technological advances.”
We’ll keep bringing you updates as this and other AI cases develop.
This article is provided for informational purposes only—it does not constitute legal advice and does not create an attorney-client relationship between the firm and the reader. Readers should consult legal counsel before taking action relating to the subject matter of this article.