Get the latest tech news
When does generative AI qualify for fair use?
Suchir Balaji 10/23/24 While generative models rarely produce outputs that are substantially similar to any of their training inputs, the process of training a generative model involves making copies of copyrighted data. If these copies are unauthorized, this could potentially be considered copyright infringement, depending on whether or not the specific use of the model qualifies as “fair use”.
And lastly they find that the average account age of a question-asker trends up after the release of ChatGPT, suggesting that newer members are either not joining or are leaving the community: This distinction -- between substituting and non-substituting uses -- is actually the origin of “fair use” from the 1841 case Folsom v. Marsh, in which the defendant copied parts of a biography of George Washington to make a version of their own. The purpose of the copying is highly transformative, the public display of text is limited, and the revelations do not provide a significant market substitute for the protected aspects of the originals.
Or read this on Hacker News
