The Distance Between a Publisher and Its Reader
Rice v. Paladin Enterprises, ChatGPT, and the line between publishing information and participating in its use
I recently came across Rice v. Paladin Enterprises (1997) in my 1L Torts class while studying aiding and abetting. As soon as I saw a publishing case in my Torts textbook, I knew I wanted to look into it more closely. The case centers on whether publishing an instruction manual not directed at any particular person could amount to aiding and abetting a battery later committed using its instructions. That made me think about the issue in publishing terms: When does putting information into the world cross the line into participating in what a reader ultimately does with it?
Rice involved a book sold to a reader the publisher did not know. Today, however, information can be delivered through systems that interact repeatedly with individual users and receive information back from them. That shift may make the line between providing information and participating in its use harder to draw.
That question feels particularly timely. Earlier this month, British Columbia sued OpenAI over its alleged failure to notify law enforcement after the shooter’s ChatGPT conversations about gun violence were flagged before the February 2026 Tumbler Ridge school shooting. The allegations have not been adjudicated, and the lawsuit involves very different facts and legal theories from Rice.
Still, the comparison highlights how much the relationship between information provider and recipient has changed. Rice considered liability where a publisher allegedly intended a harmful use despite having no specific knowledge of the reader’s criminal plan. ChatGPT, on the other hand, raises a different problem because an interactive system can continue exchanging information with the same user while drawing on the context of that conversation.
The distance between publication and use
In Rice, Paladin Press published Hit Man: A Technical Manual for Independent Contractors, which purported to instruct readers on becoming professional contract killers. James Perry purchased the book and later murdered Mildred Horn, her eight-year-old son Trevor, and Trevor’s nurse, Janice Saunders, following numerous methods described in the manual. Their families sued Paladin, arguing that the publisher had aided and abetted the murders.
Rice is not relevant to publishers just because Hit Man contained dangerous information. What made the case unusual was the manual’s alleged intent for that information to be used.
For purposes of summary judgment, Paladin stipulated that it had marketed its publications to attract and assist criminals and would-be criminals and that it intended and knew the books would be used to plan and execute murder for hire. Those concessions were limited to the motion, and Paladin reserved the right to contest them later. At the same time, the parties stipulated that Paladin had no other known contact with Perry and no specific knowledge of his murder plot.
The Fourth Circuit nevertheless rejected Paladin’s argument that the First Amendment created a blanket shield from aiding-and-abetting liability. Crucially, however, the court distinguished Rice from ordinary “copycat” cases.
A reader who imitates a crime depicted in a novel, film, article, or song has generally misused the work relative to the use its publisher intended. A crime novel can explain a murder without trying to help someone commit one; a newspaper can describe criminal methods without sharing the criminal’s purpose. The court reasoned that legitimate informational, educational, political, or entertainment purposes would ordinarily defeat an inference that a publisher intended to facilitate the crime.
That distinction protects the marketplace of ideas from a rule under which publishers would become responsible for every dangerous way someone might use what they read. Rice was exceptional because, accepting the stipulated facts, Perry’s criminal use was allegedly not a misuse of the publication at all—it was an intended use.
Rice does not eliminate the distance between a publisher and an unknown reader. Rather, it shows why that distance cannot be the entire analysis: even without specific knowledge of the eventual reader’s plans, the purpose behind the publication can still be legally significant.
What changes when the reader talks back?
Interestingly, the district court that initially dismissed the Rice lawsuit seems to have anticipated that the publisher-reader relationship would not always look so simple.
In its 1996 opinion, the court observed that “recent electronic technologies, including the Internet” were already creating complex questions about the connection between published information and later criminal conduct.
Thirty years later, ChatGPT presents a very different version of that problem.
On September 21, British Columbia filed suit against OpenAI over the February mass shooting at Tumbler Ridge Secondary School. According to Reuters, the province alleges that OpenAI’s safety team flagged the shooter’s conversations involving gun violence but law enforcement was not notified. The shooter’s account was deactivated, after which the shooter allegedly created another account and continued using ChatGPT while planning the attack. OpenAI has said that its models are trained to refuse requests that could meaningfully enable violence and that it notifies law enforcement when conversations indicate an imminent and credible threat to others.
The distinction from Rice is critical: knowing about a particular user is not the same as intending to help that user commit a crime. Rice involved extraordinary stipulations of specific intent; the British Columbia litigation instead raises questions about what an AI provider allegedly knew and what obligations, if any, followed from that knowledge.
Paladin distributed a book and had no other known contact with Perry. ChatGPT can instead respond to successive prompts from the same user, so the provider-user relationship may develop through an ongoing exchange rather than end when information is distributed.
That changes the factual distance at the center of Rice. The older case asks what happens when a publisher allegedly intends an unlawful use despite lacking specific knowledge of the individual reader; the newer dispute asks what happens when an information provider may know considerably more about a particular user without sharing that user’s intent.
The line is not simply dangerous information
That distinction is also why I do not think the lesson of Rice is that some information is simply too dangerous to publish.
Books, journalism, films, and other works routinely contain information that could be put to harmful uses. Making liability depend on that possibility would require publishers to anticipate virtually limitless reader behavior and could chill the free exchange of ideas the First Amendment protects.
Instead, Rice directs attention to the relationship between providing information and its eventual use. The Fourth Circuit focused on purpose: Was a work put into the marketplace of ideas and later misused, or was facilitating that use itself allegedly part of the reason for publishing it? The court emphasized that Rice was different from ordinary copycat cases precisely because Paladin had stipulated that the criminal use was an intended one.
ChatGPT shows why that inquiry may now require another dimension. An information provider can interact with a particular user, retain the context of that conversation, and continue responding. ChatGPT, for example, can respond to follow-up questions using context from the same chat. That does not transform knowledge into intent or automatically create liability, but it makes the traditional image of a publisher sending information into the world and then losing sight of its reader less universally applicable.
Rice suggests one answer to the question that first caught my attention in Torts: when a publisher specifically intends to assist unlawful conduct, the act of publication does not necessarily insulate it from aiding-and-abetting liability. Current litigation involving ChatGPT presents a harder and different question about what the law should make of an information provider that may know more about a particular user without sharing that user’s intent.
I went looking into Rice because I was excited to find publishing law tucked into my Torts reading. What I found was a reminder that the distance between a publisher and its reader has always mattered—and that, as technology brings information providers closer to the people on the other side, understanding what that distance protects may matter more than ever.