factnews._official on Instagram: "Research group Bellingcat discovered users exploiting Elon Musk’s Grok AI to “unblur” redacted faces of children and young women in recently released Epstein files. Between January 30 and February 5, Grok generated fabricated images in response to 27 of 31 unblurring requests, some “believable,” others “comically bad.” The disturbing trend follows Grok generating an estimated 3 million nonconsensual AI nudes in January, including 23,000 images of children. X res