Information Today, Inc. Corporate Site KMWorld CRM Media Streaming Media Faulkner Speech Technology DBTA/Unisphere
PRIVACY/COOKIES POLICY
Other ITI Websites
American Library Directory Boardwalk Empire Database Trends and Applications DestinationCRM Faulkner Information Services Fulltext Sources Online InfoToday Europe KMWorld Literary Market Place Plexus Publishing Smart Customer Service Speech Technology Streaming Media Streaming Media Europe Streaming Media Producer Unisphere Research



Vendors: For commercial reprints in print or digital form, contact LaShawn Fugate (lashawn@infotoday.com)

Magazines > Information Today > October 2023

Back Index Forward
SUBSCRIBE NOW!
Information Today
Vol. 40 No. 8 — October 2023
LEGAL ISSUES
AI and Copyright: Steering Around the Iceberg
by George H. Pike

LINKS TO THE SOURCES

“The Intelligence of Artificial Intelligence” by George H. Pike
infotoday.com/it/may23/Pike--The-Intelligence-of-Artificial-Intelligence.shtml

Register Your Work: Registration Portal
copyright.gov/registration

Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence
govinfo.gov/content/pkg/FR-2023-03-16/pdf/2023-05321.pdf

Copyright Office Announces Online Webinar Exploring Impact of Artificial Intelligence on Copyright around the World
copyright.gov/newsnet/2023/1010.html
For the May 2023 issue, I wrote my column about AI and ChatGPT. Or, more accurately, ChatGPT wrote the first third or so of the column, and I wrote the rest. I submitted it to my Information Today editor as I do with each of my column entries, all of which are covered by a writer’s contract that lays out, among other things, the copyright interests that both Information Today and I share in the piece as it is written, published, and potentially reused going forward.

There’s no doubt about the copyrightability of the portion of the column I actually wrote. Like any work that contains the “modicum of creativity” that is necessary for it to be entitled to copyright protection, the copyright in the work was established at the moment the work was created. But an extra step is often needed to ensure that a copyrighted work gets full legal protection: registering it with the U.S. Copyright Office.

COPYRIGHT REGISTRATION

The registration process for copyrights is generally easier, quicker, and less expensive than the registration processes for patents and trademarks. Completing an online form and paying a fee that can be as low as $45, along with depositing a couple of copies of the work, is usually all it takes. The normal turnaround time for approval is only 2–3 months, compared with an average of nearly 26 months for patent approval.

However, the part of my column that was written by ChatGPT is not entitled to copyright protection, and how copyright law and the registration process deal with AI material has been a difficult subject. The challenge has only increased over the last several months with the rise of ChatGPT and other generative AI applications.

HUMAN CREATIVITY

The Copyright Office is on record as saying that “copyright can only protect material that is the product of human creativity.” Previous works involving non-humans—animals that “take” photographs, for example—have been held to not be entitled to copyright protection due to their lack of a human “author.” However, the use of tools to create works does not prevent the works from being copyrighted. A camera may create the actual photograph, but it is the human photographer who makes the “original intellectual conception” that is the creativity underlying the photograph.

In March 2023, the Copyright Office issued guidance suggesting that AI tools can be similarly used to create expressive works that are eligible for copyright protection, and at a June 2023 webinar, office representatives clarified and expanded their guidance.

GUIDANCE

The March 2023 guidance statement said that the use of AI technology in creative works must be disclosed as part of the process for registering a work for legal copyright protection. The guidance further stated that “AI-generated content that is more than de minimis should be explicitly excluded from the application.” De minimis works would not be considered sufficiently creative if created by a human.

Excluding certain types of content from a copyright application isn’t particularly unusual. Copyright exists for limited times and only protects creative works. As such, there are several types of content that are not considered copyright-eligible and are excluded from copyright registration applications. These include material that has fallen into the public domain, facts, mathematical and scientific formulas, shapes and colors in graphical works, the musical scale in musical works, stock characters, and scènes à faire (i.e., standard elements in a particular genre or story, such as tropes) in books and movies. For example, it isn’t original that the Titanic hit an iceberg, it is a common creative trope that a rich girl falls in love with a poor boy, and the ship sinks to the tune of “Nearer, My God, to Thee” (a hymn by Sarah Flower Adams with a tune by Lowell Mason from the mid-1800s, which is in the public domain), so those elements can’t be copyrighted. However, the specific arrangement of those facts and scenes can be copyrighted.

‘AUTHOR CREATED’ AND ‘MATERIAL EXCLUDED’

In the process of registering a copyright, the application form requires that the applicant distinguish “Author Created” content from “Material Excluded” content. The guidance document stipulates that AI-generated material be identified as “Material Excluded,” with a brief description of the AI content. However, as noted, the arrangement of excluded material—which would now seem to include AI-created content—can be “Author Created.”

If the use of AI is de minimis, no disclosure is required as part of the registration process. Examples of de minimis include using AI as a spell or grammar checker, as a video editor, or to remove background noise from audio. In each case, there would be no need to address the use of AI.

THE PUBLIC DOMAIN

The Copyright Office suggested that one of the best ways to think about AI content is to consider it to be similar to material that is in the public domain. It is completely permissible to use public domain material either as a work in and of itself or to incorporate it in a new creative work, with the caveat that the user cannot get copyright protection for their use of the public domain work. The Copyright Office indicated that to consider AI in the same way would be both “simple and straightforward.”

While this guidance, as both originally issued in March and clarified and expanded at the June webinar, is helpful to content creators in addressing AI, it is still closer to the tip of the iceberg (to keep the Titanic references going) in dealing with AI and intellectual property issues. Copyright law does not protect ideas and has been held not to protect “styles” of art or music. AI platforms have gotten powerful enough to re-create the style of a particular music performer such as Bruce Springsteen or Taylor Swift or the artistic style of Andy Warhol, but as long as they don’t copy the lyrics, composition, or actual works, they are not likely violating copyright law, at least as of yet.

COPYRIGHT REFORM

Long before the proliferation of AI and the beginning of the AI legal debate, there have been on and off discussions in Congress and among policymakers of a need to reform the U.S. Copyright Act—last reformed in 1976, nearly 50 years ago—to address questions of fair use, the appropriate length of copyright protection, digital technology, and now AI. I’m a bit hard-pressed to see how a copyright reform effort may end, given the diverse interests and policy considerations involved. But the time is soon to come when the law may not be able to keep up.

George H. PikeGeorge H. Pike is the director of the Pritzker Legal Research Center at Northwestern University School of Law. Send your comments about this article to itletters@infotoday.com or tweet us (@ITINewsBreaks)