Journalists and advocates have been trying to grasp how many victims in total were harmed by Grok’s nudifying scandal after xAI delayed restricting outputs and app stores refused to cut off access for days.
The latest estimates show that perhaps millions were harmed in the days immediately after Elon Musk promoted Grok’s undressing feature on his own X feed by posting a pic of himself in a bikini.
Over just 11 days after Musk’s post, Grok sexualized more than 3 million images, of which 23,000 were of children, the Center for Countering Digital Hate (CCDH) estimated in research published Thursday.
That figure may be inflated, since CCDH did not analyze prompts and could not determine if images were already sexual prior to Grok’s editing. However, The New York Times shared the CCDH report alongside its own analysis, conservatively estimating that about 41 percent (1.8 million) of 4.4 million images Grok generated between December 31 and January 8 sexualized men, women, and children.
For xAI and X, the scandal brought scrutiny, but it also helped spike X engagement at a time when Meta’s rival app, Threads, has begun inching ahead of X in daily usage by mobile device users, TechCrunch reported. Without mentioning Grok, X’s head of product, Nikita Bier, celebrated the “highest engagement days on X” in an X post on January 6, just days before X finally started restricting some of Grok’s outputs for free users.
Whether or not xAI intended the Grok scandal to surge X and Grok use, that appears to be the outcome. The Times charted Grok trends and found that in the nine days prior to Musk’s post, combined, Grok was only used about 300,000 times to generate images, but after Musk’s post, “the number of images created by Grok surged to nearly 600,000 per day” on X.