AI has created a copyright headache that existing law was never designed to handle. Large language models and image generators are trained on vast quantities of human-created content - books, articles, artwork, photographs, code - often without the explicit permission of the creators. Is this fair use, or is it mass infringement? Can the outputs of an AI system be copyrighted, and if so, by whom? Courts in multiple jurisdictions are actively wrestling with these questions, and the answers are far from settled. Several high-profile lawsuits are testing whether training AI on copyrighted material constitutes infringement. The New York Times, Getty Images, and numerous individual authors and artists have brought cases against AI companies. Meanwhile, copyright offices in multiple countries have ruled that AI-generated content cannot be copyrighted because copyright requires human authorship - a position that creates uncertainty for anyone using AI to generate commercial content. For your business, the practical implications are real. If you are using AI to generate content, you should understand the legal risks around both the inputs (was the training data used lawfully?) and the outputs (can you claim ownership of what the AI produces?). These questions will be settled through a combination of court decisions and new legislation over the coming years, and the answers will shape the commercial viability of generative AI.