Threat actors have found a way to bypass link restrictions on promoted posts on X, using the platform’s Grok AI assistant. The exploit essentially tricks the AI assistant into promoting malicious links on the platform by hiding them in a From field beneath the post.