<![CDATA[Every branch of intellectual property law was built around the quiet assumption that a person sits behind the work, the invention, or the secret. Generative AI unsettles that assumption without giving courts anything clean to replace it. Ask whether an AI system can be an author, an inventor, or a party capable of keeping a promise of confidentiality, and the answer is almost always no. Still, that answer resolves very little, since it says nothing about who should be held responsible when an author is displaced, a consumer is misled, or a trade secret leaks with no one left to blame. This paper follows that gap as it shows up across copyright suits brought against AI developers in the United States, a patent dispute spanning nearly a dozen jurisdictions over whether an AI system can be named an inventor, trademark and publicity-rights claims arising from AI-generated likeness, and the quieter but arguably more urgent problem of what happens to a trade secret once it is pasted into a public chatbot. Along the way, it weighs the case-by-case litigation favored in the United States and the United Kingdom against the European Union's more pre-emptive, disclosure-driven approach. It argues that the more useful question courts and regulators should be asking is not what the AI is, but who among the people who built, deployed, or used it is best placed to answer for the harm once it occurs.]]>
Copyrights © 2026
