The author also explores potential changes to copyright law to accommodate machine learning, but deems this unlikely due to lack of political will and potential negative impacts on AI development. The author concludes that a commercial arrangement will likely be found to compensate creators while still allowing AI to learn from their works.
Key takeaways:
- The article discusses the ongoing lawsuits against AI companies like OpenAI and Github, alleging copyright infringement. However, the author believes these cases will be dismissed due to the transformative nature of machine learning.
- Copyright law protects creative works but does not extend to ideas or transformative works. This makes it difficult to argue against machine learning, which combines and transforms existing works to create something new.
- The author suggests that those suing AI companies would prefer a patent-like protection, which would allow them to block the use of their ideas even in new inventions. However, this goes against the spirit of creativity and expression that copyright law is designed to protect.
- While the author acknowledges that Congress could change copyright law to include an exception for machine learning, they believe this is unlikely due to lack of appetite for changes to copyright law and the potential negative impact on AI development.