How Human is Human? AI and Copyright
Last week the U.S. Copyright Office released Part 2 of its Copyright and Artificial Intelligence Report. We can all be forgiven for not paying close attention to this when it dropped, but if you are a content geek like me, it provides an excellent analysis of the current state of copyright and AI. And while it certainly leaves some questions unanswered (there is a part 3 pending), it starts to give clarity to a lot of questions that arts and cultural institutions may have about how the use of AI might start to more directly impact their work.
Backing up a bit, the January 29 report is the second of three reports that the Office has planned as part of a comprehensive review of how copyright and adjacent legal principles intersect with artificial intelligence.
Part 1 focused on digital replicas - that is, the use of generative artificial intelligence to create images, text, video, and other materials which look and/or sound like existing creators, be they visual artists, writers, or performers. I suspect that I’ll return to these concepts as Congress hopefully turns to addressing some of the report’s recommendations for new legislation to address these concerns.
Part 3 will be released sometime in the near future and is planned to focus on the implications of training AI models using existing works, which is another hotly debated topic and, I suspect, will include significant legislative recommendations.
With respect to copyrightability, however, the Copyright Office has concluded that existing copyright laws are sufficient to address questions regarding the degree of protection that may be extended to AI generated works.
That conclusion? Works which are created purely by AI are not protected by copyright under U.S. law.
The reality is a little more nuanced. The use of AI to organize thoughts or to further adapt an artistic prompt does not render a work non-protectable. Instead, for hybrid works, where there is a mix of human creation and AI generation, human authors are required to disclose and potentially disclaim the portions which are solely generated by AI.
There is already frustration with this conclusion, as it is not immediately clear just how much humanity has to be involved in creation to trigger copyright protection. This is not just a potential concern for multinational entertainment companies - any presenter of someone else’s artwork may need to understand exactly what rights are being conveyed in license to develop, present, perform, and adapt work that might have been created in part by AI.
And, of course, creators need to understand how to protect their rights in their own expression. If you want to experiment with using AI as part of your creative process, what do you need to do to make sure that your representation of good and valid copyright is a good and valid representation?
Before you get to some of these administrative and contractual questions, more largely, organizations may need to think about their institutional stance about the use of artificial intelligence:
As a matter of policy do you want to bar all art that might have been created using AI as part of a pro-human artist mindset?
Or do you want to encourage hybrid creation and the avenues that technology opens up for artists who might not be able to fully participate in traditional artistic mechanisms?
Do you want to encourage your staff to use AI to increase their productivity and in what areas is this permissible and not permissible?
Not all of these questions may be answerable now, but it’s time to start thinking about them and the potential legal implications of your answers. The time is ripe to consider your positions and look for the expertise that can help you find answers that your team, your board, and your artists are comfortable with.
Attorney advertising for informational purposes only; not intended as legal advice.