The misinformation spread so far that hedge fund billionaire Bill Ackman felt compelled to correct it on X. Israeli Prime Minister Benjamin Netanyahu also initially got it wrong, telling journalists it was a Jewish bystander who wrestled the gun away, The Times of Israel reported.
“Somebody invented a white guy, an ‘IT professional’ named Edward Crabtree, who stopped the Bondi shooting and spread it all over the internet, which was picked up by AI agents and slop aggregation sites,” wrote journalist Ben Collins on Bluesky.
Grok has been spouting misinformation about the Bondi massacre.Credit: xAI
Beyond the false hero narrative, Grok displayed bizarre errors throughout Sunday and Monday. When users showed the chatbot video of Ahmed tackling the shooter, it claimed the footage was “an old viral video of a man climbing a palm tree”. In another instance, it claimed a photo of the injured Ahmed was an Israeli hostage taken by Hamas.
The misinformation extended far beyond Grok.
An AI-generated image of a man having make-up and fake blood applied at the scene has flooded Reddit, WhatsApp group chats and X. The image was purported to be human rights lawyer Arsen Ostrovsky who gave interviews to media from the scene with his face covered with blood.
A cropped photo of two female police officers at the scene went viral after being posted by a “Barron Trump” fan account based in Eastern Europe, falsely claiming they “completely froze” while “gunmen reportedly fired for up to 20 minutes without interruption”. The post, which racked up more than 100,000 views, sparked a wave of misogynistic commentary attacking female officers and so-called “DEI” hiring. In fact, the uncropped image showed the officers directing civilians away from the scene after the shooters had been neutralised.
An innocent Pakistani man living in NSW told The Guardian that his life had become a “nightmare” after his photo was widely circulated falsely labelling him as one of the shooters. Naveed Akram, 30, shares the same name as one of the alleged attackers. The claim went viral after being heavily promoted by India-based accounts and was picked up by CBS News.
False claims also spread of co-ordinated attacks at Dover Heights and Double Bay, forcing NSW Police to issue a statement: “There have been NO reports of any incidents at Dover Heights – please do NOT share unconfirmed rumours.”
In another example of deliberate manipulation, someone created a fake Google Maps listing labelling the Bonnyrigg home address of one of the alleged shooters as “Bondi beach shooters place” and categorising it as a “local mosque”. The listing, which Google says has since been removed, required technical sophistication to create – involving a website with specific code to trigger Google’s business listing feature.
Ellis urged caution before amplifying unverified claims.
Loading
“Think twice before you repost things that you see or read, especially if they contain specific factual information or messages strongly biased towards triggering an emotional response,” he said. “The most reliable information tends to come out as the event is unfolding. From there, it is often swamped by misinformation – this is the danger zone we find ourselves in right now.”
Australia’s eSafety commissioner meanwhile said that her office had received multiple complaints about online material showing footage of the mass shooting, which authorities have declared a terrorist incident. The regulator reminded platforms that sensitive content labels and blurring filters should be implemented to prevent incidental exposure.
“eSafety will continue to work with platforms and services to ensure they meet their obligations under Australian law,” the regulator said. “Further actions may be considered.”
xAI was contacted for comment and responded only with its automated reply: “Legacy Media Lies.”
More Bondi terror coverage
Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.