Fair use is a doctrine in United States copyright law that allows limited use of copyrighted material without the rights holder's permission for purposes such as criticism, news reporting, teaching, and research. Section 107 of the US Copyright Act sets out four factors to weigh: the purpose and character of the use, the nature of the work, the amount used, and the effect on the market for the original. Uses that are "transformative", giving the work a new expression or purpose, are more likely to qualify.
In the AI field, the central question is whether using copyrighted books, articles, and images to train models without permission counts as fair use. AI companies argue that training is a transformative use that learns statistical patterns rather than reproducing works, while creators argue it is unauthorized use that undermines their markets. The US court decisions that found Google's book-scanning project to be fair use are a key reference point in the debate.
Many other jurisdictions have narrower exceptions, such as fair dealing in the United Kingdom, rather than an open-ended fair use doctrine.