Channeling, Translation, and the Urantia Book
Why copyright law cared about human authorship long before ChatGPT existed
Previously: ELIZA Knew You Before ChatGPT Did.
Before ChatGPT launched — before “generative AI” was a phrase anyone used at a dinner table — copyright law was already working through a question that should feel familiar: what happens when a human being claims they didn’t write the thing they wrote?
The answer the courts landed on is stranger and more useful than you’d expect. And it comes from a 2,097-page book about the cosmos, dictated by celestial spirits through a sleeping Chicago businessman, published in 1955.
A note on sources: what follows draws on a case brief I worked from rather than the primary opinions themselves. I’ve tried to flag where I’m relying on the brief’s characterization rather than reading the court’s own words. If you’re using any of this for legal purposes, go to the opinions directly.
The Sleeping Subject
Here is the setup. In the early years of the twentieth century, a Chicago physician named Dr. William Sadler began treating a patient who fell into inexplicable trances. The brief I worked from places the initial encounters somewhere between 1906 and 1911 — I can’t verify the precise date from memory. During these episodes, the patient’s vocal cords produced communications that claimed, explicitly, to originate not from him but from extraterrestrial spiritual entities on a mission to deliver new revelation to humanity.
Dr. Sadler organized a discussion group — eventually formalized as a body called the Forum, and a smaller inner circle called the Contact Commission. The Commission’s job was to pose questions to the spiritual beings, through the sleeping intermediary. The answers materialized, over years, as handwritten manuscripts. By the mid-1930s the papers were complete: 196 papers, 2,097 pages. They were published in 1955 as The Urantia Book, and the Urantia Foundation registered the copyright the following year — 1956, as the brief stated it.
The institutional position of the Foundation was clear: the book had no human author. The celestial beings wrote it. The humans were just the receiving apparatus.
This was a sincere theological claim. It was also, decades later, a legal disaster waiting to happen.
The Lawsuit
In 1990, a reader named Kristen Maaherra — who had been studying the book for over twenty years — digitized all 2,097 pages onto floppy disks and began distributing them to study groups, free of charge. The Foundation sued for copyright infringement. The brief dates the lawsuit to January 1991.
The defense Maaherra ran was simple and lethal: you said the book was written by celestial beings. The law requires a human author. You can’t have it both ways.
The district court agreed with her. As I read it through the brief, the case was Urantia Foundation v. Maaherra, 895 F. Supp. 1347 (D. Ariz. 1995). The court found that the sleeping patient, whatever his role, had functioned as a purely passive conduit — a mechanical transcriber. He claimed no creative input. He wasn’t the author. The Foundation had tried to argue the book was a “work made for hire,” which would have put the copyright in the employer’s hands, but the court found that unworkable: the Contact Commission had, by their own account, received their organizational instructions from the spiritual entities through the patient. The roles of employer and employee were inverted. The relationship was not an employment contract. It was closer to the opposite.
Without a human author, and without a valid work-for-hire claim, the district court ruled the copyright invalid. The Urantia Book entered the public domain.
The Foundation appealed.
The Ninth Circuit’s Move
On appeal, the U.S. Court of Appeals for the Ninth Circuit reversed — Urantia Foundation v. Maaherra, 114 F.3d 955 (9th Cir. 1997). The appellate court agreed with the district court on the core theological neutrality: it didn’t matter whether the celestial beings were real. Courts don’t adjudicate that. What the Ninth Circuit focused on instead was what the humans had actually done.
The Contact Commission hadn’t just sat in a room and waited. They had spent years designing, drafting, and refining specific questions to present to the spiritual entities. Those questions weren’t passive. In language the brief I read characterizes as the court’s finding, the Commission’s questions “materially contributed to the structure of the Papers, to the arrangement of the revelations in each Paper, and to the organization and order in which the Papers followed one another.”
That human-led selection, coordination, and arrangement was enough. Drawing on Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991) — the case that established that a compilation of unprotectable facts can still earn copyright if the selection and arrangement reflect originality — the court treated the divine revelations like unprotectable raw material. The humans had organized that raw material, and the copyright belonged to them.
The Foundation won on appeal. The copyright stood.
The Older Case in the Shadow
The Urantia litigation didn’t happen in a vacuum. A half-century earlier, a California court had faced almost the same question and reached the opposite result — because the facts differed in the one place that turned out to matter.
In Oliver v. Saint Germain Foundation, 41 F. Supp. 296 (S.D. Cal. 1941), a man named Frederick Spencer Oliver claimed to have transcribed a book dictated to him by a deceased Tibetan spirit named Phylos. The court found against the copyright claim. Oliver’s insistence that he had contributed nothing — that he was purely a passive amanuensis, a human dictation machine — was fatal. If you claim no creative role, the court can’t give you one.
Passive transcription, no copyright. Active compilation — questions that shaped structure, arrangement that organized the revelation — yes. The law didn’t care whether the source was celestial or subconscious or fraudulent. It cared whether a human being had made choices.
When the Law Meets the Prompt
This is where the cases start to feel less like obscure religious-IP footnotes and more like a working framework for right now.
The “AI as channel” metaphor is everywhere in legal and popular commentary on generative AI. The prompt engineer is the Contact Commission. The model is the sleeping subject. The output is the Papers. The question is whether the human who typed the instructions gets a copyright.
The U.S. Copyright Office’s 2023 guidance on AI-assisted works — the existence and general character of which is well-attested, though I’m working from secondary accounts of it rather than the primary document — draws on this lineage directly. The Office’s position, as that guidance is generally summarized, is that text prompts function more like instructions to a commissioned artist than like authorship itself, because the user can’t fully predict or control which specific creative choices the model will make. Raw AI output, under this framing, belongs to no one.
But the Office also left a path open. If a human creator selects, coordinates, and arranges AI-generated outputs — if they make the choices the Contact Commission made — the resulting compilation can be protected, provided the human-authored contributions are disclosed and the machine-generated portions are disclaimed. That characterization of the guidance’s position on compilation protection is, again, drawn from the brief rather than my own reading of the USCO document.
In 2025, the D.C. Circuit settled the hardest version of the question in Thaler v. Perlmutter, No. 23-5233 (D.C. Cir. 2025). Stephen Thaler had built an AI system he called the “Creativity Machine” and sought to register its output as a work made for hire, with the machine as the sole author. The D.C. Circuit affirmed the denial. The brief characterizes the holding as establishing that the Copyright Act of 1976 requires all eligible work to be authored in the first instance by a human being — though I’d want to read the opinion itself before treating that as a precise quotation. No human creative spark, no copyright. Thaler had claimed none.
