Meta AI tricked into handing over passwords for major Instagram accounts

In a major security breach the vulnerabilities in Meta’s rapid rollout of automated features have been exposed. Hackers have successfully used Meta’s AI-powered support chatbot to infiltrate high-profile Instagram accounts. This was aimed at targeting prominent figures ranging from Barack Obama’s White House account to Sephora and the US Space Force Chief Master Sergeant. Meta confirmed the incident on Monday (June 1), stating it had resolved the problem after independent researchers exposed the flaw.

The exploit gained traction over the weekend as everyday users reporting similar hijackings flooded Reddit and X. Security researchers and hacking groups actively circulated tutorials, videos, and screenshots on Telegram demonstrating how to bypass Meta's defenses. A video shared on X illustrated the simplicity of the attack: a hacker instructed the Meta AI assistant to link a targeted account to a completely new email address. Moving past standard security protocols, the chatbot complied, assuring the hacker that a verification code had been sent to the new email and requesting the numbers within the chat interface.

Once the hacker pasted the correct verification code, the AI assistant provided a button to reset the targeted account's password. To bypass Meta's location-based safeguards, hackers utilised virtual private networks (VPNs) to spoof the account holders' physical locations.

This incident throws light on the growing concerns regarding the safety of relying on artificial intelligence for critical security infrastructure, particularly password management. Meta has recently shifted its corporate strategy to prioritise automated systems, introducing the AI support assistant globally on Facebook and Instagram earlier this year. A March press release from Meta originally lauded the feature, stating, “The Meta AI support assistant is a major step in our work to deliver stronger support on our apps.”

According to promotional materials released during the global rollout, Meta designed the system to “take action for you on a growing set of requests directly within Facebook and in the future, on Instagram”. These automated capabilities explicitly included reporting scams, identifying impersonation accounts, flagging problematic content, and managing password resets.

Following the exposure of the vulnerability by reporting from 404 Media, Meta moved to patch the system. In an official statement addressing the exploit, Meta declared: “This issue has been resolved, and we are securing impacted accounts.” However, the company has not yet disclosed the exact number of users affected by the breach.