Many people have asked me about AI and copyright law in the past few months. It is a hot topic right now with the introduction of ChatGPT, and there are several possible uses for AI within the entertainment industry. At the same time, AI technology raises a number of issues that the courts and legislatures are still addressing. This article looks at three big issues facing AI technology, which will probably need to be resolved before the technology can truly take off.
The first issue to be addressed is the copyrightability of AI-generated content. The Copyright Office has definitively stated that content must have a human author in order to have a copyright; they will not register anything authored by a machine (or a monkey, as in Naruto v. Slater, a.k.a. the “monkey selfie” case). So anything that is 100% AI generated is in the public domain and free of copyright protection. But what if an author uses AI as part of their creative process? The Copyright Office has suggested that a computer can be used as an aid to a human author. Still, the question remains whether the human author can claim full copyright protection or if the copyright would be limited to his contribution, not the AI-generated part. While this question has not yet been answered, I suspect that the amount each contributed will be an essential factor in the final decision. Another important factor will likely be the amount of creative input the human gave to the AI program; some countries are allowing copyright protection when the directions show sufficient creative input.
The second issue to be addressed is AI using copyrighted content to generate its own content. In order for AI to do what it does, it must have an extensive database of content to reference, and that database does include copyrighted content. Currently, most AI algorithms do not include limitations on using copyrighted content, so it is very easy to instruct AI to create something that incorporates iconic characters or text. While this certainly raises a concern about AI being used for evil, the bigger problem would be people using AI-generated content and not realizing that that content contains copyrighted material. Knowing that AI-generated content might contain copyrighted material could very easily put a damper on people using it. On the flip side, owners of lucrative copyrights are kicking up a fuss over their content being in the database without their permission. Some are suing the owners of the AI content, while others are figuring out how to work with the owners to prevent their content from being incorporated into AI-generated works without permission. At this point, it’s hard to say whether this will be resolved by the courts, the legislature, or by the various parties working it out amongst themselves.
The third issue is how AI will alter employment opportunities for content creators. AI is a faster and cheaper way to create content, which could significantly impact the job market for animators, special effects experts, and writers, to name a few. Content creators must figure out how to compete with this new kid on the block. In addition, Author’s Guild recently put out a press release warning writers that their content could be used for AI training purposes and recommended that future contracts include a clause expressly prohibiting this (https://authorsguild.org/news/model-clause-prohibiting-ai-training/). It’s a good point that paid writers should take to heart – who wants to find out they’ve unknowingly been training their replacement?
Time will tell how these issues, and the others raised by AI, will ultimately be resolved. In the meantime, if you have questions about how AI and copyright law can be used for your benefit or detriment, feel free to contact me at firstname.lastname@example.org.