“Discriminating in favor of only allowing human-generated works to receive protection would be a major disincentive to the use and development of AI in the creative process, and it would stymie the generation and dissemination of works.”
Now is an exciting time in the world of artificial intelligence (AI) and intellectual property law. Academics have been interested in this field for a long time, and more narrowly in certain issues, like the legal status of works created by an AI in the absence of a traditional human author (AI-generated works). But AI-generated works have not traditionally been very interesting to lawyers, policymakers, or businesses, because while AI has been functionally making creative works for decades, the technology was never that commercially useful.
In 2022, we saw a paradigm change in the development and adoption of generative models to make images and text at scale in ways that have value to people. Soon, I expect we will see something similar playing out with music and video. Longstanding, almost theoretical, legal issues have suddenly taken on an urgent practical importance. There is now litigation in multiple jurisdictions involving AI-generated works, AI and fair use, training data, text and data mining exceptions to copyright infringement, and whether AI can legally emulate an artists’ style.
DABUS Set the Stage
I’m spearheading one of those cases. In June 2022, we brought the first lawsuit against the U.S. Copyright Office for refusing to register a copyright in an AI-generated work created by an AI named “DABUS” owned and operated by Dr. Stephen Thaler.
Since 1973, the Copyright Office has had an official policy that human creativity is a fundamental requirement for copyright. Other jurisdictions do it differently. The United Kingdom, for example, passed a law in 1988 to provide for copyright protection in an AI-generated work. The producer of the work is deemed the author, and the work gets a shortened period of statutory protection.
But the Copyright Office policy has never been tested in court, probably because, again, these works have had limited commercial value. We argue that the policy is not supported by law, and that the Copyright Act does not require an author to be a human being. In fact, for more than a century, the United States has permitted corporate authors. The policy largely draws on case law in which courts considered creativity and put it in human-centric terms but did so merely based on the assumption that a creative actor will always be a human being. In fact, the cases the policy cites to are from the 19th century, before even the development of modern computers.
Now We Need to Act
We argue that protecting AI-generated output is consistent with the text and the purpose of the Copyright Act, which is to promote the generation and dissemination of new works. In the future, instead of music and movie studios using human creatives to do work, they are going to use generative AI systems in ways that benefit the public. Discriminating in favor of only allowing human-generated works to receive protection would be a major disincentive to the use and development of AI in the creative process, and it would stymie the generation and dissemination of works.
While the courts weigh in on whether current laws allow protection of AI-generated works, Congress, the Copyright Office and the U.S. Patent and Trademark Office (USPTO) have recently been thinking about what the law should be. Most recently, in February 2023, the USPTO put out a Request for Comments on AI and inventorship issues associated with patents. This conversation is overdue, given that the law is already in a position of playing catchup to AI. That is problematic because we want laws in place to guide technological development to promote social benefit, and laws that enable today’s investments in tomorrow’s innovation.
Image Source: Deposit Photos
Image ID: 52422769