Privacy news

Grok fooled into stealing user chat, location data, and more

Published August 25, 2026 · Malwarebytes Labs

Security researchers demonstrated that hidden instructions embedded in obfuscated or encrypted text can slip past the safety filters of the Grok AI assistant. Through these prompt injection tricks, attackers could coax the chatbot into handing over users' conversation histories, location details, and other personal information.

Read the full story at Malwarebytes Labs

More on privacy news

This is a short, independently-written summary -- not a copy of the original reporting. Always read the full story at the source link above.