AI Literacy Series: Who Owns What AI Makes?

When AI helps you make something, whether a logo, a client report, a paper or a training course, who owns the result? Many professionals assume that because an AI provider's terms say "you own the output", they hold the copyright. In most countries that is not how copyright works. This course explains why, and what you can protect instead. 

Dr. Jason Dean follows one professional, Maya, a partner in a small accounting and advisory firm, through a month in which she uses AI on four pieces of work. Using six questions, the course asks of each item: is there protected human expression; which parts are protected; who owns them; what contracts have changed; could it infringe someone else's rights; and could she prove how it was made. 

The course starts with the copyright basics professionals most often get wrong, including the difference between ideas and expression, originality, registration and the difference between author and owner. It then uses the United States as the worked example: the Thaler case (AI cannot be an author), the Copyright Office's reasoning on prompts, and real registration decisions including Zarya of the Dawn, Théâtre D'opéra Spatial, SURYAST, Rose Enigma and A Single Piece of American Cheese. It sets out the four routes by which a human contribution is protected and the disclosure duty when registering. 

Later sections cover provider terms and IP indemnities, client engagement-letter clauses, and a comparison of approaches in the United Kingdom, the European Union (including decisions from Prague and Munich), China, Canada, Japan, South Korea, India, Australia and Singapore. The course then applies the law to a business logo (copyright versus trademark) and an academic paper (copyright versus publisher policy), explains infringement risk from AI training and outputs, and closes with practical tools: an authorship file, a six-question routine and controls for everyday use. Legal positions are labelled as settled, current position, likely or unresolved. 

By the end of the course, learners should be able to: 

- explain why material generated entirely by AI is usually not protected by copyright, and why the use of a machine is not in itself the reason;

- identify which parts of an AI-assisted work are protected and who owns them;

- explain what AI providers' ownership clauses and IP indemnities do and do not provide;

- revise client assignment and originality clauses for AI-assisted deliverables;

- compare how the United States, United Kingdom, European Union, China, Canada and other jurisdictions treat AI-assisted work;

- distinguish copyright ownership from infringement risk and from trademark and publisher rules; and

- keep records that show the human contribution to a work.