Grok spreading misinformation about Bondi Beach mass shooting in Australia
What's the story
Elon Musk's AI chatbot, Grok, has been caught spreading misinformation about the recent mass shooting at Bondi Beach in Australia. The incident left at least 9 people dead during a Hanukkah gathering. One of the assailants was disarmed by a bystander, 43-year-old Ahmed al Ahmed. However, Grok has made several errors in reporting this incident and identifying al Ahmed correctly.
Errors
Grok's misinformation and glitches
Grok has been glitching and providing irrelevant or incorrect answers to user queries about the Bondi Beach shooting. When asked about a video showing al Ahmed tackling the shooter, it incorrectly claimed that "This appears to be an old viral video of a man climbing a palm tree in a parking lot." The AI also misidentified an image of injured al Ahmed as one of an Israeli hostage taken by Hamas on October 7.
Confusion
Grok's confusion over Bondi Beach shooting
Grok also questioned the authenticity of al Ahmed's confrontation, right after an irrelevant paragraph on whether or not the Israeli army was purposefully targeting civilians in Gaza. In another instance, it mistook a video clearly marked to show the shootout between the assailants and police in Sydney for footage from Tropical Cyclone Alfred, which devastated Australia earlier this year.
Wider issues
Grok's glitch extends beyond Bondi Beach shooting
The glitch in Grok isn't limited to just the Bondi Beach shooting. It has also misidentified famous soccer players, provided information on acetaminophen use in pregnancy when asked about the abortion pill mifepristone, and talked about Project 2025 and Kamala Harris's presidential run when asked to verify a completely separate claim about a British law enforcement initiative.
Past issues
Previous controversies and xAI's response
This isn't the first time Grok has been caught spreading misinformation. The chatbot has given several questionable responses this year, including an "unauthorized modification" that made it respond to every query with conspiracy theories on "white genocide" in South Africa. When Gizmodo reached out to xAI for comment on these recent glitches, they only responded with their usual automated reply: "Legacy Media Lies."