First-Hand:Computer Art Copyright
Submitted by A. Michael Noll
August 26, 2023
© 2023 AMN
AI seems everywhere – including the creation of art. What is the role – if any – of a human, and can AI-art be copyrighted and registered at the US Copyright Office?[1] The question of copyright registration of computer art was illuminated and resolved in 1965, and this story is recounted here.
In April 1965, the Howard Wise Gallery in New York City exhibited computer-generated patterns by Bela Julesz and A. Michael Noll, who were employed as researchers at Bell Telephone Laboratories, Inc. in Murray Hill, NJ. Julesz’s patterns were used for investigations of stereovision; Noll’s patterns were early digital computer art. All the works shown were copyrighted in the names of the creator: Julesz or Noll.
All that was required to copyright a work was to place the copyright symbol ©, date, and name of the artist on the work. The date was the date of public exhibition or publication – not necessarily the date of the work’s creation. In the 1960s, exhibition or publication was a requirement for copyright.
After the Wise exhibition, it was suggested that I register the copyright with the US Copyright Office. I decided to register the copyright of my work “Gaussian-Quadratic.” It was programmed in FORTAN on an IBM 7090 mainframe computer and plotted on 35 mm microfilm by a Stromberg Carlson SC-4020 printer. Although “Gaussian-Quadratic” was created around 1963/4, the copyright date was the public exhibition date of 1965.
The computer program and algorithm combined a quadratic equation with Gaussian randomness to create the image. Computers do nothing at random, and the “randomness” consisted of numbers that appeared random to humans and were created by a program.
I filled in the copyright form and posted it to the Copyright Office, describing the work as computer art. The Office replied that a copyright could not be registered for a work done by a machine. I then described that a program by a human controlled the machine, and that the program combined a mathematical equation with elements of randomness. The Office then replied that a work produced by randomness could not be registered. I then explained that a program calculated the “randomness”, and the numbers produced only appeared “random” to humans – algorithms did all. The Copyright Office was then finally satisfied, and the copyright was registered.[2]
If it is believed that with AI the computer is the creator, then human copyright registration is not allowed. But if “AI” is ultimately just algorithms programmed by a human, then copyright registration should be possible. But the AI advocates and promoters would probably not want to diminish the hype by admitting any human involvement.
