Three developments landed within days of each other, and together they define the US position on AI copyright as clearly as it’s been stated yet.
The U.S. Copyright Office published Part 2 of its “Copyright and Artificial Intelligence” report on March 17, 2026. The conclusion is direct: works created entirely by AI cannot receive copyright protection under current US law. Human authorship is required. The report does leave a door open for human-AI collaborative works, but the threshold matters. The Copyright Office’s standard is that the human author’s creative contribution must be substantial. Not present. Substantial. That word carries legal weight, and it should be treated accordingly by anyone documenting AI-assisted creative workflows.
The Supreme Court reinforced the same conclusion procedurally. Earlier this month, on March 2, the Court denied certiorari in Thaler v. Perlmutter (No. 25-449). To be precise about what that means: the Court did not rule on whether AI can hold copyright. It declined to hear the case, which leaves in place the lower court ruling that requires human authorship. The practical effect is the same – AI-only ownership claims are foreclosed at the federal appellate level, but the legal mechanism is procedural, not a merits ruling.
Then came the lawsuit. Encyclopedia Britannica and Merriam-Webster filed suit against OpenAI in mid-March 2026. Reuters reported that Britannica is seeking unspecified monetary damages and a court order blocking the alleged infringement. The complaint goes further than training data alone. The Verge’s coverage indicates the lawsuit alleges ChatGPT-based products reproduce near-identical copies of Britannica and Merriam-Webster content, cannibalizing traffic directly. This is active litigation, it alleges, it doesn’t prove. But the output reproduction claim has structural implications that training data disputes don’t fully capture.
The convergence is the story. Policy guidance, a closed judicial door, and active litigation all arrived in the same reporting window. For companies using generative AI to produce content, the question is no longer whether the US will take a position, it’s whether their current workflows can document sufficient human contribution to meet the standard the Copyright Office has now set.