So far, public debate in the Czech Republic has mostly produced views on whether content created by AI (images, photographs, text, legal opinions and so on) is or is not likely to be protected as a copyright work. I do not yet see a blanket answer.
What interests me is a question one step further back. Does the AI itself infringe copyright when it compiles, abstracts and otherwise processes works protected by copyright — including the author’s own rights, such as the right of reproduction and distribution? Litigation on exactly this point is already under way abroad against the developers of AI (Microsoft, GitHub and OpenAI).
A distinction has to be drawn between so-called “transformative AI”, where a photograph is simply taken and modified by the AI with various filters, and so-called “generative AI”, usually understood as a model that creates from scratch rather than merely modifying an existing copyright work. Even there, the question of a potential infringement of the author’s rights is not ruled out — an AI will hardly ever create entirely “from nothing”.
