Apple warned Grok over deepfakes

Apple privately warned xAI in January that Grok could be kicked out of the App Store if it kept generating sexualized deepfakes of real people. (nbcnews.com) The warning came in a letter Apple sent to three United States senators after they pressed the company about Grok’s image tools and whether the app was violating App Store rules. (nbcnews.com) Apple told senators it had already required xAI to make changes, then said the first round of fixes was not enough and demanded stronger limits on the app. (9to5mac.com) xAI responded by restricting who could use Grok’s image-generation features and by limiting edits involving real people, steps that kept the app in Apple’s store. (macrumors.com) Apple’s App Store rules say the store is meant to provide a “safe experience,” and the company says apps must keep changing to stay on the platform. (developer.apple.com) The episode lands as xAI is pushing Grok deeper into government work. FedScoop reported this week that xAI is pursuing Federal Risk and Authorization Management Program High authorization, with the United States Department of Agriculture sponsoring the effort. (fedscoop.com) Federal Risk and Authorization Management Program, or FedRAMP, is the federal government’s cloud-security review system, and its artificial intelligence track now prioritizes conversational tools meant for routine use by federal workers. (fedramp.gov) xAI already has a foothold in Washington. In September 2025, the General Services Administration announced a OneGov deal that made Grok models available to federal agencies for $0.42 per organization for 18 months, through March 2027. (gsa.gov) That leaves xAI managing two tests at once: satisfying Apple’s consumer-platform safety rules while trying to meet the federal government’s security and procurement standards. (developer.apple.com) (fedscoop.com)