OpenAI flagged a Florida man’s violent ChatGPT threats to the Federal Bureau of Investigation (FBI), and agents moved in after he vowed to kill his ex by month’s end.
Story Highlights
- Reporting says OpenAI alerted the FBI in May 2026 after detecting explicit murder threats.
- Court records quoted the man saying, “I’m gonna kill her by the end of this month”.
- The case ties a named suspect and Palm Beach proceeding to the ChatGPT messages.
- OpenAI’s public safety language says it notifies police for imminent, credible threats.
OpenAI Alert Triggers FBI Action in Florida Threat Case
News accounts identify a 25-year-old Florida man who used ChatGPT to spell out rape and murder plans against his ex-girlfriend. Reporters say OpenAI detected the threats and sent an alert to the Federal Bureau of Investigation (FBI) in May 2026, which drew law enforcement into the case. The quoted chat content was direct and time-bound. According to reporting on court-language, the man wrote, “I’m gonna kill her by the end of this month,” and, “If I can’t have her then nobody can”.
Coverage links the threats to a real court proceeding in Palm Beach County. Articles name Darren Zhou, describe him as a former financial analyst, and report a local case that resulted in probation after threats against his ex-girlfriend. Those details root the story in a specific venue and outcome, not a rumor. While outlets did not publish the underlying docket or FBI intake form, the reporting across pieces presents the same chain: threat, OpenAI alert, FBI involvement, court result.
What OpenAI Says About Reporting Imminent Harm
OpenAI’s public materials say the company notifies law enforcement when conversations indicate an “imminent and credible risk of harm to others”. The company also explains that it uses automated systems and human review to monitor activity for safety, in line with its policies. That broader policy context matches what reporters describe here: a violent, time-bound threat that crosses the line from offensive speech into an explicit plan to do harm, prompting a referral to authorities.
Legal and policy experts have tracked how chat logs from artificial intelligence tools are now turning up as evidence. Analysts note that these chats are not protected like talks with a lawyer, doctor, or therapist, so platforms can retain them and disclose them in defined cases, including emergencies or with legal process. As more people use chatbots for sensitive topics, these records become part of the same trail that texts, emails, and social media already provide to police and courts.
Why This Matters for Safety, Privacy, and Free People
This case shows how threats online can reach real victims fast, and how a private platform may step in before someone gets hurt. Most readers will see a clear line here: a named man allegedly promised to kill his ex by a date, and police got a heads-up. That fits common-sense policing. At the same time, it reminds us that companies are watching for violent threats, and that your chats are not a vault. People should treat these tools like any public forum when it comes to serious risks.
Florida man told ChatGPT he'd murder his ex. OpenAI alerted the FBI https://t.co/Z288fKKMGF via @pbpost
— Nick Penzenstadler (@npenzenstadler) August 14, 2026
For conservatives, two pillars apply. First, protect life and punish crime. Threats like this deserve swift action, and victims deserve peace. Second, guard civil liberty and limited government. Americans should know how and when companies escalate chats to police, and what standards they use. OpenAI says it reports imminent, credible harm. That line should stay tight and clear. Strong rules focus on real danger, not political speech or lawful gun rights, faith, or family values.
Sources:
youtube.com, yahoo.com, reuters.com, wsj.com, politico.com
