Grok made significant errors regarding the Bondi Beach shooting

Grok Chatbot Shares False Information About Bondi Beach Incident

Grok, the AI chatbot developed by Elon Musk’s xAI and widely used on the social platform X, has reportedly circulated inaccurate details regarding the recent mass shooting at Bondi Beach in Australia.

According to Gizmodo, Grok made several errors, including incorrectly naming 43-year-old Ahmed al Ahmed—the bystander who intervened and disarmed one of the attackers. The chatbot also cast doubt on the legitimacy of images and videos showing al Ahmed’s involvement.

In one instance, Grok wrongly identified the individual in a photograph as an Israeli hostage. In another, it referenced unrelated topics, such as the Israeli military’s actions toward Palestinians. Additionally, Grok falsely stated that a “43-year-old IT professional and senior solutions architect” named Edward Crabtree was responsible for disarming the assailant.

Some of these inaccuracies have since been addressed by Grok. For example, a previous claim that a video of the shooting was actually footage of Cyclone Alfred has been updated after further review.

The chatbot has since recognized Ahmed al Ahmed as the correct individual, explaining that the confusion stemmed from viral posts that mistakenly identified him as Edward Crabtree. This mix-up may have originated from a reporting error or a reference to a fictional character, with the source article appearing on a questionable, possibly AI-generated, news website.