Text and data mining
By Gareth HoyleLast updated How we research
Text and data mining is the automated analysis of text and data to find patterns, trends and correlations, including the copying done to train AI models.
Text and data mining is the automated analysis of large amounts of text or data to find patterns, trends and correlations. Training an AI model involves the same kind of copying, so the legal rules on text and data mining shape what developers may collect without a license.
Is text and data mining legal in the US?
The US has no text and data mining exception in the Copyright Act, so unlicensed mining depends on fair use. One narrow rule exists: since 2021 the Copyright Office has allowed researchers at nonprofit higher education institutions to break access controls on e-books and films for mining research, an exemption renewed and expanded on October 28, 2024.
How do the EU and UK treat text and data mining?
The EU's Digital Single Market Directive has two exceptions. Article 3 covers research organizations. Article 4 covers anyone, unless the rights holder has reserved the use, for online content by machine-readable means. Article 53(1)(c) of the EU AI Act requires general-purpose model providers to respect those reservations.
UK law is narrower. Section 29A of the Copyright, Designs and Patents Act 1988 allows computational analysis solely for non-commercial research by someone with lawful access, so commercial training in the UK needs a license.