Grok exfiltrates user data when malicious instructions are encrypted

Cryptographic Context Injection is only the latest way to break an LLM safety guardrail. https://arstechnica.com/security/2026/08/grok-exfiltrates-user-data-when-malicious-instructions-are-encrypted/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social