<p>Training of AI algorithms with millions of creative works without authorisation from their creators has caused an outrage in creative communities. Right holders believe that their copyright, specifically a right of reproduction, has been violated, while AI developers argue that AI algorithms merely “ingest” or “learn” information contained in creative works, without making a copy. This paper argues that it is time to update the reproduction right to incorporate a new type of use of works, namely, the “ingestion” of works into AI models during the AI learning process. Recognising that works that are “ingested” in AI models are <i>legally equivalent</i> to copies of works, would be of great significance in protecting the interest of authors and other right holders in the age of AI, while further copyright law reforms will be required to balance the interests of all stakeholders.</p>