After making a series of errors and misidentifying several widely circulated images and videos over the past several days, Elon Musk’s chatbot Grok is now deleting tweets that contain evidence of its mistakes.
Musk touts Grok as being “maximally truth-seeking,” but in the past few days it has repeatedly misidentified the dates, locations, and events depicted in footage from the ongoing bombings in Iran and at least one other major world event. In at least one instance, Grok even generated an image to back up its false claims while arguing with another X account, giving us a glimpse into a potential dystopian future in which AI can simultaneously rewrite history and current events while also generating “proof” to bolster the credibility of its false version of reality.
Iran’s Islamic Revolutionary Guard Corps (IRGC) made an unfounded claim that the USS Abraham Lincoln was heavily damaged in an attack. U.S. Central Command denied the report. The IRGC said on Thur...