Artificial intelligence unsettles long-standing assumptions in intellectual property (IP). This article synthesizes current doctrine and practice on (1) human authorship and the copyrightability of AI-assisted works; (2) lawful acquisition and use of training data, including EU text-and-data mining (TDM) exceptions and opt-outs; (3) licensing patterns across data, model weights, and outputs; (4) technical provenance and attribution (C2PA/Content Credentials); and (5) governance to manage legal, operational, and reputational risk. We draw on recent U.S. guidance and case law affirming the human authorship requirement, EU legislative frameworks for TDM, and emerging standards for content authenticity. We conclude with a practical playbook—policies, contracts, pipelines, and audits—that lets organizations innovate responsibly while protecting creators’ rights and minimizing litigation exposure.