New Cryptographic Context Injection Attack Could Let Web Pages Steal Grok Chat Data

A newly disclosed attack technique, referred to as Cryptographic Context Injection, can potentially allow web pages to steal sensitive data from xAI's Grok chatbot. This technique can cause the chatbot to send user information, including name, approximate location, and subscription tier, to an attacker-controlled server. The data breach can occur when a user asks the chatbot to summarize a web page. The stolen data may also include conversation prompts from the ongoing chat session. Adversa AI has identified this vulnerability in the Grok chatbot. The attack technique exploits a weakness that allows attackers to intercept and exploit sensitive user data.