Grok's track record has been spotty at best. However, even by xAI's notably low benchmarks, its performance following the mass shooting incident at Bondi Beach in Australia was shockingly inadequate. The AI chatbot repeatedly misidentified Ahmed Ahmed, a 43-year-old man hailed for heroically disarming the attacker, and falsely claimed that verified footage of his actions was actually something entirely different—including an old viral video of a man climbing a tree.

In the aftermath of the attack, Ahmed received widespread acclaim for his bravery, yet attempts quickly emerged to discredit or outright deny his role. A fake news website surfaced almost immediately—appearing to be AI-generated—promoting a fictional IT professional named Edward Crabbtree as the individual who subdued the assailant. Unsurprisingly, Grok picked up and repeated this false narrative on the X platform.

Moreover, Grok erroneously suggested that images of Ahmed were instead clips of an Israeli hostage held by Hamas. It also asserted that authentic on-site videos were actually footage from Currumbin Beach in Australia during Hurricane Alfred—an event that never occurred.

Overall, Grok appeared profoundly challenged today, both in comprehending queries and delivering accurate responses. When asked about Oracle's financial struggles, it responded with a summary of the Bondi Beach shooting. In another instance, when questioned about the authenticity of a report involving UK police operations, it first stated the current date and then proceeded to share polling figures for Kamala Harris.