AI and Copyright: Who’s That Author Behind the Computer?

October 11, 2019

Morgan Lewis intellectual property partner Ron Dreben recently hosted an entertaining and informative session for the firm’s global intellectual property practice, with updates on artificial intelligence and copyright legal issues, and his predictions for the future. Here are some key takeaways from the session:

Companies will likely want to invest in, develop, and protect AI developed content. What can clients do to ensure that AI is viewed as a tool used in creating a work?

  • Ensure that your company owns the AI system software code you develop or have developed, and that software is registered at the US Copyright Office. When applicable, try to protect the source code or AI algorithms as a trade secret. Any protection efforts should be done in coordination with patent counsel.
  • The infringement risk of using copyrightable works for AI training should be considered early in the process. Based on recent changes to its copyright law, Japan may be a favorable country for the creation of machine learning databases.
  • To the extent an employer or investor wants a company to own the AI output, it is important to have a human employee (or contractor who has assigned all relevant rights) be part of the creative process, so that a human can reasonably be identified as an “author” in a copyright application.
  • Machines will increasingly get better at producing creative works, further blurring the lines between artwork that is made by a human and by an AI system. Legislation will likely be needed to address ownership of some AI output and whether copying for commercial machine learning purposes is fair use in the United States.
  • Enforcement activity may in some cases be easier outside the United States. Commencing AI-related infringement action outside the United States may be more likely to lead to a worldwide settlement.
  • More synthetic media will be created and distributed. Harmful deepfakes should be reported to the FBI. The distribution of harmful deepfakes may constitute criminal copyright infringement, among other civil and criminal violations. In some cases, fair use may protect deepfakes. We can handle these issues by combining our industry regulatory practices with our copyright experience.
  • AI developed content will likely have many IP, economic, social, and even national security implications.

If you have any questions or would like more information on the topics above, please contact Ron Dreben.