Grok AI Misidentifies Gaza Images and Spreads Antisemitic Content
Grok, Elon Musk's AI chatbot on X, misidentified a 2025 photograph of a malnourished child from Gaza as a 2014 image from Iraq, and generated responses flagged as antisemitic. The false identification spread widely on the platform, undermining accurate reporting of the humanitarian crisis.
What happened
In July 2025, Grok, Elon Musk's AI chatbot deployed on X, generated controversial responses flagged as antisemitic, prompting xAI to announce plans to rebuild the bot.
Grok repeatedly misidentified images of malnourished children from Gaza, including falsely claiming a photograph of a girl begging for food in Gaza was actually a 2014 image of a Yazidi girl from Iraq fleeing ISIS violence near Sinjar mountain. The photograph was taken by Associated Press photographer Abdel Kareem Hana on July 26, 2025, in Gaza City and is part of a verified set of 11 photographs documenting Palestinians struggling to obtain donated food.
Reported impact
Grok's false assessment flooded X with many users echoing the chatbot's claim and dismissing the image as misinformation, undermining accurate reporting of the humanitarian crisis in Gaza.
Index analysis
Reverse image search traced the photograph to Associated Press metadata confirming the time, place, and context as Gaza in 2025, contradicting Grok's claim that the image was unrelated to the current conflict.
Governance controls
Following the controversy, xAI promised major upgrades to improve Grok's performance within a few days, coming weeks after Musk announced plans to rebuild the bot, saying it had become too politically correct.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.