AI-Generated Content & IP Rights: Humans Required?
In the context of Intellectual Property, ownership of traditional creations is relatively straightforward: a patent vests with the inventor under 35 U.S.C. § 100(f), a copyright vests with the author (or employer in case it’s a work made for hire) and trade secrets belong to the individual or entity taking necessary steps to protect them. However, when it comes to Artificial Intelligence, this concept of ownership becomes more complex.
Standard rules usually determine ownership of AI-related IP, such as patented AI systems, copyrighted source code, or trade secret data sets. However, disputes may arise about ownership of inventions, software, and other valuable data created by AI itself instead of a natural person or legal entity. For instance, if AI were to develop an
improved apparatus or self-generated computer code, or generates valuable data sets.
As established in the Patent Act and by court decisions, an inventor must be a natural person--not a corporation or artificial intelligence system. Specifically, the patent act requires an inventor to be an “individual.” 35 U.S.C. § 100(f). The Federal Circuit (who oversees patent appeals) has held that inventors must be natural persons, as opposed to corporations. Beech Aircraft Corp. V. EDO Corp., 990 F.2d 1237, 1248 (Fed. Cir. 1993). The US District Court for the Eastern District of Virginia has held that an inventor must be a natural person (see Thaler v. Hirshfield, 2021 WL 3934803 (E.D. Va. Sep. 2, 2021). And the U.S. Patent and Trademark Office’s Manual of Patent Examining Procedure (MPEP) references a "person" with regard to inventorship. MPEP § 2137.01.
For more on AI’s current inventive capability (or at least as of 2018), please see Section II(B) in my previous publication in the Texas Law Review titled The USPTO’s Sisyphean Plan: Increasing Manpower Will Not Match Artificial Intelligence’s Inventive Capabilities.
In a further update, a recently issued report from the USPTO addressed intellectual property issues raised by AI. The USPTO sought public comment on whether “a work produced by an AI algorithm or process, without the involvement of a natural person . . . qualif[ies] as a work of authorship” under the Copyright Act. U.S. PATENT AND TRADEMARK OFFICE, PUBLIC VIEWS ON ARTIFICIAL INTELLIGENCE AND INTELLECTUAL PROPERTY POLICY at 19 (2020). In its summary of responses, USPTO noted that “the vast majority of commenters acknowledged that existing law does not permit a non-human to be an author [and] this should remain the law.” Id. at 20–21.
The Copyright Act does not clearly define the term "author," but courts and the US Copyright Office have established that authors must be human. This point was asserted in the "monkey-selfie" case, where the Court of Appeals for the Ninth Circuit held that a monkey did not have copyright rights to photographs it took of itself. See Naruto v. Slater, 888 F.3d 418 (9th. Cir. 2018). As a result, the Copyright Office also clarified that any work of authorship must be created by a human. See Copyright Review Board’s Decision (Feb. 14, 2022). Furthermore, the work made for hire doctrine generally does not cover commissioned works of computer software.
The Copyright Office has recently refused to register the copyright claim in an AI-generated work. In the application, the work’s creation was describe being “autonomously created by a computer algorithm running on a machine.” The Copyright Office refused the registration because it “lacks the human authorship necessary to support a copyright claim.” See Copyright Review Board’s Decision (Feb. 14, 2022). Although, the Copyright Review Board was not able to consider whether the Work was created with contribution as the applicant argued the human authorship requirement is unconstitutional and unsupported by case law.
Consequently, there was no analysis as to the degree of human involvement necessary in the creation of machine-generated works in order for the work to meet the statutory criteria for copyright protection. As a basis of support, the Copyright Review Board cited the Compendium Section 313.2, which in turn quotes a Copyright Office Report from 1966! Specifically, it states “the ‘crucial question’ of human authorship is whether a computer is ‘merely being an assisting instrument’ or ‘actually conceive[s] and execute[s]’ the ‘traditional elements of authorship in the work’”). Ultimately, the Board found that Congress limited Copyright protection to human authorship but left the door open for Congress to change that.
In the context of trade secrets, ownership of AI-created technology is relatively insignificant as the entity that takes formal steps to protect their trade secret retains respective rights to resultant technology and data. Conclusion & Take-Aways Intellectual property laws and regulations that determine ownership of AI-generated creations are more complex than those for traditional creations. Patents and copyrights must be held by natural persons, not corporations or artificial intelligence systems. Thus far, it appears that there is a general consensus between government and non-government entities that only Congress will be able to change this, however, the judicial system should be closely watched as novel and creative arguments are likely to surface around these complex ownership questions in the event of IP infringement and/or licensing disputes.