The interaction between copyright and artificial intelligence is a new area, and many of the legal issues have yet to be decided by Congress or the courts.
Under current U.S. law, only human can hold copyright; Ais cannot. But what about a work created by a human using AI to create a portion of the work? Is that portion copyright protected because it is part of the grater whole?
Devices like ChatGPT depend on having a large pool of text images or sounds, from which they learn to produce new material. If the material in the data set is copyrighted, is that use infringement or fair use?
The New York Times is currently suing Open AI (the creators of ChatGPT) and Microsoft for copyright infringement over just this issue. They claim that their copyrighted material is being unfairly used to compete with them. For more information, see" "NY Times sues OpenAI, Microsoft for infringing copyrighted works", Reuters, December 27, 2023. Individual authors have also sued OpenAI for misuse of their copyrighted works to train its AI tools.
Elsevier has added language to its subscritption licenses forbidding mass downloading of their content for AI training without permission. The publisher has also developed policies on the use of ChatGPT and other generative AI by authors, and reviewers who work with Elsevier.