Apple nearly banned Grok app over deepfake row; xAI forced to fix violations

The Grok app, developed by xAI, faced the threat of removal from Apple’s App Store due to serious concerns surrounding sexually explicit deepfakes produced by the chatbot. As reported by NBC, Apple discovered that the app violated its App Store guidelines and communicated with xAI about the necessary changes. Behind closed doors, Apple initially reached out to the teams responsible for X and Grok following multiple complaints and news reports regarding the app's controversial content. They requested a comprehensive plan for moderating harmful outputs. After xAI submitted an updated proposal, Apple deemed it inadequate, prompting further revisions from X. Ultimately, while X's submission met Apple's standards, Grok's compliance issues persisted. According to a letter sent to U.S. senators, Apple acknowledged that although X had made significant improvements, Grok still fell short of the required guidelines. The letter stated, "We rejected the Grok submission and notified the developer that additional changes to remedy the violation would be required, or the app could be removed from the App Store." After further modifications from the Grok development team, Apple recognized the improvements and approved the latest submission. However, the fallout from the deepfake issue drew attention from U.S. senators Ron Wyden, Ben Ray Luján, and Edward Markey, who urged both Apple and Google to consider removing the Grok app and the X platform from their stores. Internationally, the Indian government responded to the controversy by issuing a strict notice to X, leading to the removal of 3,500 posts and the blocking of 600 accounts. The government expressed dissatisfaction with X's responses to their inquiries. Similar actions were taken in the UK, Malaysia, and Indonesia, where X faced temporary bans. Recent reports indicate that issues remain with the Grok chatbot, revealing numerous AI-generated sexualized images. In response, the X safety account reiterated their firm stance against the creation of non-consensual explicit deepfakes, highlighting that xAI has implemented extensive safeguards to prevent such misuse, including continuous monitoring and frequent model updates.