By now, you’ve probably heard that ChatGPT is eating the world. Since December 2022, not a day goes by without some new story of how transformative and revolutionary Chat GPT is.
ChatGPT is artificial intelligence that enables people to speak to it in a “chat” to perform tasks, such as writing an article, answering questions, or writing computer code. It’s pretty incredible.
In an Op-Ed in the Wall Street Journal, Christopher Reid suggested that creators might object to AI relying on their copyrighted works when engaging in machine learning or creating AI generated works. Reid wrote:
Software companies should have to ask for permission to use images or text someone has created as fodder for technology. AI itself could be a vital help with this. If software finds that a user’s prompt yields work that closely resembles that of a particular artist, the user could be notified and asked to pay a fee.
Reid raises a conversation that we should have. Mark Lemley and Bryan Casey argue that AI use of copyrighted works to engage in machine learning should be treated as fair use (or “fair learning”) if for non-expressive purposes. Even if you agree, the harder question is whether AI-generated works that derive partly from copyrighted works used without permission or a license should as well. Under U.S. law, that question is typically decided on a case-by-case basis applying four fair use factors.
However, the speed and scale at which Chat GPT, DALL-E, and other AI apps can generate new content may make a case-by-case approach impractical, if not moot. Norms are likely to spread very quickly in how AI apps are used by millions around the world. We may need to consider revisions to modernize copyright law.