This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Google's Gemini and X's Grok AI systems provided false information about the authenticity of war imagery from Iran, incorrectly claiming authentic photographs were fake and from different locations and times.
During the US-Israeli war on Iran, Google's Gemini AI and X's Grok AI assistant provided inaccurate fact-checking information about authentic war imagery. When users queried these AI systems about photographs showing mass graves in Minab, Iran, both systems incorrectly claimed the images were fake. Gemini stated the photo was from Turkey's 2023 earthquake, while Grok claimed it was from Indonesia's 2021 COVID burials. Both AI systems provided authoritative-sounding responses with fabricated sources and non-existent links. The cemetery image was actually authentic, as confirmed by researchers who cross-referenced it with satellite images and video footage. This incident is part of a broader pattern where AI-generated misinformation is proliferating during the Iran war, with factcheckers reporting that nearly half of viral falsehoods are now AI-generated. A 2025 international study found that about 50% of AI-generated summaries had significant sourcing or accuracy issues, rising to 76% for Google's Gemini interface. The false information wastes investigative time and risks enabling atrocity denial.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
