The Supreme Court yesterday declined to hear a case brought by a computer scientist whose “invention” was in fact created by artificial intelligence. Stephen Thaler was appealing a Federal Circuit decision that interpreted the Patent Act to require a human “inventor” for purposes of obtaining a patent. The invention at issue was conceived of by Thaler’s AI model DABUS and not by a human, dooming its chances of obtaining patent protection.
The law of inventorship is now quite murky. Professor Dennis Crouch wrote yesterday about the ethical issues presented to patent attorneys in the wake of the decision. If AI-created inventions are not patentable, then is a patent attorney required to interrogate their client to determine which parts of the invention were human versus AI-created?
Of course, the elephant in the room is this: what inventions will be protectible when every R&D operation utilizes artificial intelligence to get ahead? Extending to copyright law, where Thaler is fighting a similar battle, isn’t there a “modicum of creativity” leading to the AI-created work? History suggests the law will split inventorship down the middle and grant rights to those “portions” of an invention or copyrightable work that were created by a human, and avoid granting protection to the “portions” that were AI-generated.
Copyright law does this well with the protection of derivative works. The term “Derivative Work” is defined in 17 U.S.C. § 101:
A “derivative work” is a work based upon one or more preexisting works…
Examples of creative works include a translation of a novel into another language. Or a screenplay adaptation of a novel. Or—remember in the 90s when rappers would sample 80s music? The end song would be a derivative work of the 80s tune. The “author” of the end work (translation, screenplay, 90s rap song) owns the copyright on the derivative work. But the author must have a license to the original copyrighted work, or the original work must not be copyrighted (e.g., an expired copyright). In this manner, copyright law separates the original creative expression from the preexisting copyrighted work. The same can be done for AI-generated works, with the human expression protected to the extent possible, and the AI-generated expression excluded from protection.
Patent law has also separated protectible and unprotectible portions of the inventions. For example, 35 U.S.C. § 101 controls which inventions are eligible for patent protection. As early as 1852, the law precluded the patentability of abstract ideas that would effectively preempt the abstract idea. Le Roy v. Tatham, 55 U.S. (14 How.) 156, 175 (1852) (“A principle, in the abstract, is a fundamental truth; an original cause; a motive; these cannot be patented, as no one can claim in either of them an exclusive right.”). More recently, the Supreme Court maintained that an abstract idea was not patentable, but a practical application of an abstract idea may be worthy of patent protection. Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 573 U.S. 208, 216 (2014) (“[I]n applying the §101 exception, we must distinguish between patents that claim the ‘buildin[g] block[s]’ of human ingenuity and those that integrate the building blocks into something more”). Look for patent law to handle AI inventions in much the same way and to continue carving out patent protection of AI-generated inventions through the lens of §101. What was created by AI cannot be protected and will be carved out, much the way anything other than a “practical application” of an abstract idea cannot be patented. But the elements of human ingenuity giving rise to the invention will surely be protected to continue the incentive for technological innovation.