Epstein Files: X Users Are Asking Grok to ‘Unblur’ Photos of Children

Bellingcat reports that X users repeatedly prompted Grok to “unblur” redacted faces in images from the latest Epstein file releases, producing fabricated or misleading AI-generated outputs that sometimes went viral; following inquiries, Grok appeared to stop attempting such unredactions. The article contextualizes this misuse with other incidents of Grok-enabled false identifications and deepfakes, outlines investigations by multiple governments, and reviews X’s evolving—and criticized—guardrails that first limited image tools to paid users and later further restricted editing of real people.