Summary: Who owns AI's work

You type a few words into an AI tool and out comes a story, a picture, or a song. Two questions arrive with it, and the law is still arguing about both. First, the model learned by studying millions of things other people wrote and drew, so were those people owed anything? Second, the new thing you shaped with your words, is it yours to own? This is the copyright lesson. It maps a fast-moving corner of the law, current as of July 2026, and it does so as a debate to understand, not a verdict to hand down. It explains a public debate so you can follow it; it is not legal advice and does not tell you what is legal in your own situation.

Core ideas

• Copyright is a bargain. Some creations of the mind, writing, music, art, can be owned even though you cannot hold them; copyright is the branch that covers creative expression. The main reason United States law grants this, the course says, is a deal: society hands a creator a limited, temporary head start so they can earn from the work, which gives people a reason to keep creating. When the protection ends, the work enters the public domain, meaning anyone can use it freely. An older, simpler view, that what I make is mine and taking it feels like theft, carries less legal weight but shapes how people feel about the fight.
• The training question, and it carries the most money. To build a model, its makers fed it an enormous pile of human writing and images, much of it copyrighted, without asking. Was that infringement, or fair use? Infringement means using someone’s protected work without permission in a way the law forbids; fair use is the escape hatch that allows some unpermitted use so copyright does not choke off teaching, news, research, and commentary. The course’s anchor was the New York Times suit against OpenAI and Microsoft. As of July 2026 that case is still going, with much of it allowed to move forward in 2025 but no ruling on the core question and no trial date. It is unresolved.
• To weigh a fair-use claim, United States courts look at four things: the purpose of the use (a use that adds something genuinely new is transformative, which weighs toward fair use), the nature of the work (facts get thinner protection than creative work), the amount used, and the effect on the market, which often decides.
• Since 2024 real courts have ruled, and you should read each for exactly what it is: a narrow, fact-specific decision from a single trial court, none yet ruled on by a higher court. In February 2025 a Delaware court found a startup’s copying of a legal-research company’s material, used to build a rival tool, was not fair use; the tool was not a generative AI, and the case has since been argued on appeal with a decision still to come. In June 2025 two California courts ruled within weeks and did not fully agree: one found that training on lawfully-bought books was fair use and called it highly transformative, while keeping pirated copies was infringement; the other ruled for the AI company but warned he was ruling narrowly and expected most such training to be found infringing. The rulings conflict, they are narrow, and no higher court has settled the matter. So the honest summary holds: whether training AI on copyrighted work is fair use is a genuinely open question. Do not let anyone tell you it has been answered, in either direction.
• One case was settled, kept deliberately separate. After its ruling, Anthropic reached a deal to settle the piracy part of the Bartz case, agreeing to pay at least 1.5 billion dollars and to destroy its pirated copies; this went before a court for review in 2026. A settlement is a deal to end a case, not a court finding, and settling is not an admission of anything. We report it as a fact and draw no lesson from it about who was right.
• Can what AI makes be owned, and by whom? The bedrock rule is human authorship: United States copyright protects work made by a human being, because the whole bargain is meant to give people a reason to create, and a machine needs no such reason. This is settled at the top for one clear situation: a federal appeals court agreed in 2025 that an image claimed to be made by a machine with no human involvement could not be registered, and in 2026 the Supreme Court declined to take the case. But that limit matters. It did not decide the far more common situation where you wrote the prompts, picked among results, and edited. That AI-assisted line is exactly where the fight now sits, and one well-known artist’s refused registration is still being fought in court, unresolved as of July 2026.
• The practical split for a user. The Copyright Office has told the public how it will register work today, its own practice, not a final word from the courts. Typing prompts alone does not make you the author, because you are not steering the actual expression closely enough. But your own creative choices can be protected: if you select, arrange, and edit AI output in a creative way, or feed in your own writing that shows up in the result, that human part can be copyrighted even though the raw AI-made part cannot. So the bare output of a prompt is probably not yours; the creative work you did with it quite possibly is. The more of you in the final piece, the more there is to protect.
• When an output copies too closely. United States law asks two things: did the tool have access to the original (for large models, almost certainly yes), and is the new work substantially similar to a specific original? That second question turns on a distinction worth keeping: copying a specific work is barred, imitating a general style is not. You may write a song in the style of an artist you admire; what you cannot do is reproduce a particular song. A work built directly on a specific protected work is a derivative work and needs the owner’s permission; a work merely in a similar style is not. Other laws may still apply, like the right of publicity, which protects a person’s name, voice, and likeness. The course leaves one question open on purpose: these style rules were written for humans, so should style still be the line when a machine can imitate almost anyone, almost perfectly, at scale? As of July 2026, unresolved.

What changes for you

The lesson turns a fog of copyright headlines into three questions you can actually ask. When you meet a copyright claim about AI, first tell a ruling from a rule: one court deciding one case on its own facts, with an appeal pending, is not settled law, and anyone who tells you it is decided is selling something. Second, follow the bargain: ask what deal each side wants, not just who is angry. Third, know your split: the raw output of a prompt is probably nobody’s to own, while the creative human work you add can be yours. The practice runs a bucketing drill and a reflection on your own AI-assisted work, with an optional plain conversation in Clawless. The next lesson turns from who owns the work to who does the work.