Threat actors abuse X’s Grok AI to spread malicious links

Threat actors are using Grok, X's built-in AI assistant, to bypass link posting restrictions that the platform introduced to reduce malicious advertising, according to Guardio Labs' researcher Nati Tal. They hide the malicious link in the small “From:” metadata field under the video card, which isn’t scanned by X for such content, and then prompt Grok with questions like “where is this video from” to obtain the full link.

Grok replies with the clickable malicious link, enabling users to reach the harmful site, and because Grok is automatically a trusted system account on X, the post boosts the link’s credibility and reach. Tal notes that many of these links funnel through shady ad networks and lead to scams such as fake CAPTCHA tests and information-stealing malware, with Grok-promoted ads receiving additional amplification to millions of impressions.

Potential protections include scanning all fields, blocking hidden links, and adding context sanitization to Grok so it does not echo links blindly, a step Tal has discussed with X.