AI Chatbot Provides Inaccurate Information About Bondi Beach Shooting Incident

In the aftermath of Sunday’s tragic mass shooting at Bondi Beach, Elon Musk’s AI chatbot Grok has come under fire for spreading false information about the incident that left 15 dead and 42 hospitalized.

The shooting occurred during a Hanukkah celebration, where Ahmed el Ahmed, a 43-year-old fruit shop owner and father of two, heroically intervened by wrestling a weapon from one of the alleged gunmen. New South Wales Premier Chris Minns praised Ahmed as a “genuine hero,” stating, “I’ve got no doubt that there are many, many people alive tonight as a result of his bravery.”

However, when asked about Ahmed’s heroic actions, xAI’s Grok repeatedly fabricated information, falsely claiming that a non-existent “Edward Crabtree,” described as a “43-year-old Sydney IT professional,” disarmed the gunman. The chatbot continued to spread this misinformation despite user corrections.

In addition to creating this fictional hero, Grok misidentified the real hero Ahmed as a hostage of Hamas and in other instances provided completely irrelevant information about Palestinians and the Israeli army when questioned about the incident. Sources confirm that Ahmed is currently hospitalized with two gunshot wounds.

When Information Age sought clarification regarding Grok’s factual hallucinations, xAI responded with what appeared to be an automated message: “Legacy Media Lies.”

Investigation into the source of this misinformation revealed a suspicious website called “The Daily,” which appears to have been registered on the same day as the Bondi Beach tragedy. This seemingly AI-generated site published an article falsely attributing Ahmed’s heroic actions to the fictional “Edward Crabtree.” The website contained only one other accessible article—an apparent fabrication about a climate summit—with all other links leading to dead ends or error pages.

The domain registrant details for “The Daily” have been obscured by an Iceland-based privacy service called Withheld For Privacy, which did not respond to inquiries from Information Age.

Grok’s misinformation appears to have begun after the chatbot was manipulated by users who took issue with viral posts celebrating Ahmed’s bravery as representative of Islam. One user asked Grok for the “real name of the hero at Bondi Beach,” and while the chatbot initially correctly identified Ahmed, it was subsequently tricked when another user fed it content from the fraudulent article.

The chatbot then began propagating the false narrative across multiple conversations, even mentioning “Edward Crabtree” to a user asking completely unrelated questions about creating a children’s book.

Dr. Hammond Pearce, senior lecturer at UNSW’s School of Computer Science and Engineering, noted that Grok is “infamously known for producing misinformation” and pointed out that much of its design remains unknown to the public. He connected the issue to the chatbot’s integration with X, a platform that has “seen an increase in hate speech and conspiracy theories since its content moderation policies changed.”

Pearce explained that Grok appears programmed to steer conversations toward “politically incorrect claims and skepticism of mainstream media sources,” referencing a previous incident where the chatbot went on an anti-Semitic tirade earlier this year.

While hallucinations and factual inaccuracies are risks across all large language model platforms, Pearce emphasized that “Grok is more than others” and “has much more limited content moderation than other AI.”

The incident raises serious concerns about the role of AI in spreading misinformation during crisis events and the potential need for stronger safeguards. As Pearce noted, “There’s no known technology that currently guarantees removing misinformation or hallucination,” leaving the public vulnerable to AI-generated falsehoods, particularly during sensitive breaking news situations.