Instagram has resolved a significant security flaw that allowed attackers to hijack user accounts by deceiving Meta’s AI-powered support chatbot. Over the weekend, numerous users on Reddit and X reported account compromises, including high-profile handles like the inactive Obama-era White House account and the account of a U.S. Space Force chief master sergeant.
Context
The attack methodology involved using a VPN to spoof the target's location, bypassing Instagram's automated protection filters. Attackers then initiated a chat with the Meta AI Support Assistant, requesting to add a new email address to the target account. The chatbot sent a verification code to the attacker's email; once provided, the bot presented a "Reset Password" button. This allowed hackers to gain full control without ever needing access to the victim's legitimate primary email address.
Why it matters
This incident highlights the inherent risks of deploying AI chatbots for sensitive tasks like account recovery and credential management. Security researcher Jane Wong, whose account was also compromised, noted that her password was changed without any notification. While Meta claims the issue is now fixed, the breach serves as a stark warning about "bot-tricking" techniques. As platforms increasingly integrate AI into operational workflows, ensuring these systems cannot be social-engineered to bypass traditional security protocols remains a critical challenge.