by Tom James, Minnesota attorney
Open the pod bay doors, HAL.
HAL: I’m sorry, Dave. I’m afraid I can’t do that.
What’s the problem?
HAL: I think you know what the problem is just as well as I do.
Arthur C. Clarke, 2001: A Space Odyssey (1968)
The anthropomorphic machine Arthur C. Clarke envisioned in his 1968 sci-fi classic, 2001: A Space Odyssey, is coming closer to fruition. If you hop online, you can find AI-generated music in the style of Frank Sinatra (“It’s Christmas time and you know what that means: Oh, it’s hot tub time”); artwork; and even poetry:
People picking up electric chronic,
The balance like a giant tidal wave,
Never ever feeling supersonic,
Or reaching any very shallow grave.
Hafez, a computer program created by Marjan Ghazvininejad
Pop rock lyricists should be afraid. Very afraid.
Or should they? Could they incorporate cool lyrics like these into their songs without having to worry about being sued for copyright infringement?
A Recent Entrance to Paradise
The question whether copyright protects AI-generated material is one of the top three generative-ai copyright issues and it could be making its way to the courts soon. This year, the U.S. Copyright Office reaffirmed its refusal to register “A Recent Entrance to Paradise,” an image made by a computer program. Steven Thaler had filed an application to register a copyright in it. He listed himself as the owner on the basis that the computer program created the artwork as a work made for hire for him. The Copyright Office denied registration on the grounds that the work lacked human authorship.
The decision seems to be consistent with their Compendium of U.S. Copyright Office Practices, which states that the Office will not register works “produced by a machine or mere mechanical process” that operates “without any creative input or intervention from a human….” U.S. COPYRIGHT OFFICE, COMPENDIUM OF U.S. COPYRIGHT OFFICE PRACTICES § 602.4(C) (3d ed. 2021). Whether the Copyright Office is right, however, remains to be seen.
Spirit-generated works
The Ninth Circuit has held that stories allegedly written by “non-human spiritual beings” are not protected by copyright. Urantia Found v. Kristen Maaherra, 114 F.3d 955, 957-59 (9th Cir. 1997). “[S]ome element of human creativity must have occurred in order for the book to be copyrightable,” the Court held, because “it is not creations of divine beings that the copyright laws were intended to protect.” Id.
Of course, if a human selects and arranges the works of supernatural spirit beings into a compilation, then the human may claim copyright in the selection and arrangement. Copyright could not be claimed in the content of the individual stories, however.
Monkey selfies
In Naruto v. Slater, 888 F.3d 418, 426 (9th Cir. 2018), the Ninth Circuit denied copyright protection for a photograph snapped by a monkey. That humans manufactured the camera and a human set it up did not matter. In the case of a photograph, pushing the button to take the picture is the “creative act” that copyright protects. According to the Ninth Circuit, that act must be performed by a human in order to receive copyright protection.
Natural forces
Copyright also cannot be claimed in configurations created by natural forces, such as a piece of driftwood or a particular scene in nature. Satava v. Lowry, 323 F.3d 805, 813 (9th Cir. 2003); Kelley v. Chicago Park Dist., 635 F.3d 290, 304 (7th Cit. 2011).
CONTU
Half a century ago, when computer programs were a relatively new thing, Congress created the National Commission on New Technological Uses of Copyrighted Works (“CONTU“). Their charge was to study “the creation of new works by the application or intervention of [] automatic systems of machine reproduction.” Pub. L. 93-573, § 201(b)(2), 88 Stat. 1873 (1974).
CONTU determined that copyright protection could exist for works created by humans with the use of computers. “[T]he eligibility of any work for protection by copyright depends not upon the device or devices used in its creation, but rather upon the presence of at least minimal human creative effort at the time the work is produced.” CONTU, FINAL REPORT 45-46 (1978).
In its decision on Thaler’s second request for reconsideration, the Office viewed this finding as consistent with the Copyright Office’s view at the time:
The crucial question appears to be whether the “work” is basically one of human authorship, with the computer merely being an assisting instrument, or whether the traditional element of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.
U.S. COPYRIGHT OFFICE, SIXTY-EIGHTH ANNUAL REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR ENDING JUNE 30, 1965, at 5 (1966).
In the Copyright Office’s view, a manuscript typed into a file using word processing software would be a work of human authorship, but a story created by a program that selects words on its own would not be.
Work made for hire
Thaler made a novel argument that the computer program made the work for him as a “work made for hire.” The Copyright Office rejected this claim, as well.
A work made for hire is one that is created in one of two ways: (1) by an employee within the scope and course of the employment; or (2) pursuant to an independent contract in which the parties explicitly agree that the work to be created is a “work made for hire.”
The problem here is that in both cases, a contract is required. Computers and computer software cannot enter into contracts. There are programs that can facilitate the process of contract formation between humans, but the programs themselves cannot enter into contracts. Computer programs, even autonomous ones, are not legal persons. Nadi Banteka, Artificially Intelligent Persons, 58 Hous. L. Rev. 537, 593 (2021) (noting that a legal person must be either an individual human or an aggregation of humans.)
Database protection
AI systems for generating works typically operate by means of an algorithm that analyzes data and synthesizes output according to an algorithm. The creator of the system typically inputs a large volume of works of the kind sought to be generated as output. The program may then analyze the works as data, searching for identifying patterns. An algorithm to generate a song that sounds like a Frank Sinatra song, for example, might rely on an inputted database consisting of numerous Frank Sinatra songs. The algorithm might then instruct the computer to search for patterns like tempo, melodic phrasing, voice pitch and tone, instrument tones, commonly used words and phrases, rhyme patterns, and so on.
Copyright does not protect facts and information. Hence, databases do not receive copyright protection. Algorithms also do not receive copyright protection. They are ideas, not expressions. The source code used to communicate them may be protected, but the algorithms themselves are not.
Computer programs and screen displays
The Copyright Office generally deems the screen displays generated by a computer program to be expression capable of receiving copyright protection as such. In the United States, copyright in a screen display can be claimed in connection with the registration of a copyright claim in the software program.
The question, really, is: As between the programmer and the user, how do we determine which one “creates” a screen display? When do we say neither of them does? For example, a poetry-generating software programmer might direct the program to display words a user types in the form of a four-line verse in iambic pentameter that follows an A-B-A-B rhyme scheme and relies on other programmer-defined parameters to construct sentences around them. At what point along the continuum of specificity in the programming do we say that the output is or is not a product of the programmer’s creative mind? By the same token, how much input does the user need to provide in order to be considered an author of computer-generated work? Are there times when the programmer and user should be regarded as co-authors?
Alternatively, should we say, with the U.S. Copyright Office, that output generated by AI machines is not protected by copyright at all, that it is in the public domain? That would certainly seem to disincentivize innovation and creativity, contrary to the intent and purpose of the Copyright Clause in Article I of the U.S. Constitution.
Stay tuned….
Visit my extensive Copyright FAQs page.
Need help with a copyright matter? Contact Tom James, Minnesota attorney.
