Musk's Grok accurately predicted when US would attack Iran
What's the story
In a shocking turn of events, Elon Musk's AI chatbot, Grok, accurately predicted the date of the recent US strike on Iran. The prediction was made during a methodological test conducted by The Jerusalem Post on February 25. It came just days before the actual attack took place on February 28. The test involved four AI chatbots: Claude, Gemini, ChatGPT, and Grok, each asked to pick a single day for a hypothetical US strike on Iran.
Unique forecast
How Grok outperformed other chatbots
While Claude and Gemini were unable to provide a specific date for the attack, OpenAI's ChatGPT gave a tentative date of March 1. However, it later revised its prediction to March 3. In contrast, Grok was the only one that provided a clear date in its original response ie. February 28. The chatbot said this was based on its 'most informed, evidence-based prediction,' with some speculating it might be linked to the outcome of Geneva talks.
Endorsement
Musk backs Grok amid geopolitical tensions
Musk has been a vocal supporter of Grok, often highlighting its advanced capabilities. After The Jerusalem Post article was published, he tweeted his approval by saying, "Prediction of the future is the best measure of intelligence." This is seen as an endorsement for Grok's predictive abilities amid ongoing geopolitical tensions.
AI challenges
Grok's stellar run and controversies
Grok had a stellar run late last year when it briefly dethroned Gemini and other AI tools to top several AI benchmarks. However, that reign was short-lived as Gemini 3 Pro models soon reclaimed the crown. Despite its impressive performance, Grok has also faced criticism from civil society and governments over its alleged role in creating sexualized deepfakes of women and children.