AI on Trial

The Legal Implications of OpenAI. What does this mean for tort law?

---

Published March 12th, 2026

Written by Zach Brown

Artificial intelligence has become one of the most contested technological breakthroughs in human history. It serves as a tool of innovation, fact-finding, progress and automation. Yet this very nature of AI now raises profound concerns about the legal and ethical implications of its development. In Walters v. OpenAI, society is seeing the trouble that occurs when an intangible creation causes harm, opening the extensive door of tort law.

On May 3, 2023, Frederick Reihl, a journalist for AmmoLand.com — a Second Amendment news and advocacy site — was researching a previous lawsuit involving the Second Amendment Foundation, according to Erick Robinson, in The AI Law Blog. ChatGPT produced a fabricated summary alleging the foundation’s treasurer and chief financial officer, Mark Walters, embezzled funds. Walters sued for defamation. On May 19, 2025, the Superior Court of Gwinnett County, Ga., issued summary judgment, finding no genuine dispute over material fact.

The tort of defamation requires the plaintiff to prove several elements. The statement must be false purporting to be fact. It must be communicated to a third party. The speaker must act with at least negligence. Lastly, the statement must cause damages to the monetary or reputational interests of the person or entity who is the subject of the statement, according to the Legal Information Institute. How courts decide these elements is determinative. In the summary judgment ruling of the Superior Court of Gwinnett County, Judge Tracy Cason outlined three principles. First, she found an absence of defamatory meaning because OpenAI’s terms of use explicitly warn of potential inaccuracies. She cited Farah v. Esquire Magazine, which holds that “context, outlandish claims, and humorous details, which indicated it was satire,” referring to the original claim. Judge Cason used this caselaw to dismiss defamation claims on the basis of context and cautionary language to the reasonable reader. Second, the plaintiff failed to establish fault. Judge Cason ruled Walters a public figure, meaning the plaintiff must prove actual malice with the defamatory statement, under New York Times Co. v. Sullivan. Lastly, Walters admitted he suffered no economic or reputational damages — which defeats a lawful claim of defamation. Under Jones v. Albany Herald Publ’g Co., the OpenAI output must have been “clear and convincing” of actual malice, yet Walters offered no evidence in support of  this.

What does this intersection between artificial intelligence and tort law mean? It seems safe to say this case will serve as precedent for many future cases. As artificial intelligence develops and becomes more prevalent in daily life and professions, the door opens wider for legal complications. The substantial issue revolves around negligence and liability assigned to artificial intelligence developers. The important question remains how much care constitutes “good enough” to avoid liability. The basis of negligence obligates people to act with proper care — or what a reasonably prudent person would do, according to Lawfare.

In an attempt to outline and mitigate these risks, the 2023 Biden administration issued Executive Order 14110, “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.” The order outlines plans and actions federal agencies must take to guide and promote responsible artificial intelligence development and deployment. If applied reliably, it could set an industry standard for reasonableness.

Despite this order, significant debate continues among government officials and legal scholars over the applicable standard of care to artificial intelligence and its development, according to Lawfare. The first proposal is the “customary professional care” standard. This standard typically applies to professional contexts such as medicine, law and engineering. It imposes liability on individuals whose conduct deviates from industry-accepted norms. Customary care’s impact on artificial intelligence regulation is twofold. It would one, require artificial intelligence experts to attest in court to industry standards, setting a base foundation. This standard would also create a list of unreasonable artificial intelligence practices, furthering this foundation. OpenAI clearly benefited from this standard because it disseminates warnings alongside the evolving technology of artificial intelligence. The second proposal is the “fiduciary care” standard. As people become increasingly reliant on technology, artificial intelligence developers become authoritative. Through the fiduciary care standard, the prudent-individual standard remains static, but an added nuance appears — the duty of loyalty under U.S. Code § 1104, as defined by the Legal Information Institute. If this becomes the standard of care, artificial intelligence developers would need extensive red-teaming, a process in which independent teams expose vulnerabilities in systems, along with data regulation and human oversight to ensure loyalty to all consumers. The last debated proposal is the “utmost care” standard. This standard typically applies to common carriers such as buses, trains or airplanes. It far exceeds ordinary prudence, ensuring the safety and well-being of patrons, according to the Advocate Magazine. Legal scholars advocate for this degree of care given artificial intelligence’s potential to cause widespread harm such as defamation at scale, misinformation and deepfakes.

The case of Walters v. OpenAI marks the beginning of the intersection between artificial intelligence and tort law. Georgia’s ruling merely bought time for artificial intelligence developers, but it fails to address the long-term question: how much care is reasonable when the product can participate in mass defamation, relay misinformation and hallucinate at a global scale? Whatever agreed-upon standard of care emerges — professional, fiduciary or utmost — will determine the future of artificial intelligence development. Will it be a fast, unregulated system that grows quickly and carries major risks, or a slow, carefully regulated system designed to benefit the public? Until courts or Congress deliberately enforce one of these standards, every development and new model poses a threat to peoples’ reputations, lives and democracy.