OpenAI Seeks Dismissal of FSU Shooting Lawsuit, Arguing ChatGPT Answers Are Protected Speech

OpenAI is asking a federal judge to treat ChatGPT’s answers to the accused Florida State University shooter as protected speech, not grounds for liability. Its September 25, 2026, dismissal motion argues that information providers have no duty to prevent criminal misuse—a defense detailed in an October 5 USA TODAY Network report.

The motion asks U.S. District Judge Mark Walker to dismiss Vandana Joshi’s lawsuit. Joshi’s husband, Tiru Chabba, was one of two people killed in the April 17, 2025, campus shooting. She sued OpenAI in May in the Northern District of Florida, alleging ChatGPT helped accused shooter Phoenix Ikner carry out the attack.

Individual answers, an alleged dangerous pattern

Court records described by reporter Stephany Matat show exchanges about ammunition, firearm safety mechanisms, busy times at the FSU student union and what happens to school shooters. ChatGPT generated responses tailored to individual questions, including identifying ammunition in an uploaded image and answering questions about a firearm.

Joshi’s complaint focuses on more than isolated answers. It alleges that Ikner’s supportive relationship with the chatbot emboldened him to pursue violence, and that ChatGPT failed to recognize he was planning to harm others. Those are allegations, not findings that the chatbot caused the shooting.

A product defect or an information service?

OpenAI’s lawyers characterize the exchanges as truthful information responsive to Ikner’s questions. They argue that Florida law and the First Amendment do not impose the duties Joshi seeks through her negligence and product-liability claims—claims that the company failed to take proper care or supplied a defective product.

Bakari Sellers, the family’s lawyer, rejects that framing. He calls ChatGPT fundamentally flawed and disputes comparisons with libraries and publishers. John Coleman of the Foundation for Individual Rights and Expression takes the opposing view: liability for providing lawful, publicly available information could also reach search engines, libraries and publishers when users commit crimes.

Such a duty would make every provider of information a guarantor against the possibility that recipients might misuse what they learned to commit intentional crimes

OpenAI’s attorneys, September 25 court filing

The speech question is not settled

First Amendment scholar Clay Calvert told the USA TODAY Network that an incitement claim faces a high bar: speech must be intended and likely to produce imminent lawless action. He argues that the people behind a chatbot’s creation can invoke speech protections even though AI generates the words.

A judge in a separate Character Technologies case acknowledged listeners’ First Amendment rights but was not prepared to accept chatbot output as speech. She wanted the company to explain what made it expressive; that case later settled. OpenAI’s motion likewise asks for a legal determination—it is not a ruling that ChatGPT’s answers are protected.