AI and Copyright: Who Owns AI-Generated Content?

AI and Copyright: Who Owns AI-Generated Content?

When we were building Pactum, we faced a straightforward question: could we use a logo created by Midjourney in our commercial product? It turned out the answer depends on three factors: which AI service you're using, how creatively you directed it, and where you plan to use the result. Let me break it down.

Quick takeaways:

• Under Uzbek law, only humans can be authors; pure AI output without creative human input isn't protected by copyright.
• If you wrote a detailed prompt, edited the result, or integrated generation into your product—you have arguments for ownership; if you just clicked a button—it's a gray area.
• The service license (OpenAI, Midjourney, Stable Diffusion) determines whether you can use the output commercially; always read the Terms of Service.
• Training models on third-party data carries separate risks: original copyright holders may file claims (lawsuits are already happening in the West).

What the Law Says: Authorship Requires a Human

Copyright law in Uzbekistan (like in most countries) protects works created by human creative effort. An algorithm itself cannot be a legal subject. This means:

• If an AI autonomously generated text or an image, formally the output doesn't fall under classical copyright protection.
• If a human made substantial creative contributions (crafted a complex prompt, selected one from thousands of variants, processed it in an editor)—there's a basis to consider the final work a result of collaborative effort, with the human as author.

In practice, there's no case law on this topic in Uzbekistan yet; we rely on basic principles of the Law