The public domain is the status of works and data that intellectual property laws do not protect. People can generally use them without asking permission or compensating the original copyright owner. section 9.3.3
Works and data can enter the public domain in several ways:
- Copyright can expire because a work is old. section 9.3.3
- Some material is not copyrightable, including factual information, statistics, and data generated without substantial human creative input. section 9.3.3
- Law can place work in the public domain; US federal work products are one stated example. section 9.3.3
- Public-domain material may be combined with data released under terms such as the MIT license or particular Creative Commons licenses. section 9.3.3
Why public-domain data is not automatically good training data
A large collection of public-domain works may be dominated by older texts. Those texts can contain outdated scientific information, unfamiliar language, stereotypes, and cultural assumptions that differ sharply from current audiences. Public-domain status removes some copyright concerns, but it does not remove quality, relevance, bias, or legal-status questions. section 9.3.3 section 9.6
