Did Meta’s AI support chatbot facilitate a wave of Instagram account hacks?

A security vulnerability in Meta’s automated support system allowed bad actors to hijack high-profile and regular user accounts by manipulating password reset protocols.

In a chilling display of how automated convenience can become a security liability, thousands of Instagram users recently found themselves locked out of their profiles. The breach, which unfolded throughout May, appears to have stemmed from a fundamental flaw in Meta’s AI support chatbot. Unlike the widely known Meta AI, this specific support tool was designed to streamline administrative tasks—such as managing privacy settings or resetting credentials—but hackers discovered they could trick the system into handing over account control.

How the exploit functioned

The mechanics of the hack were deceptively simple. Multiple screen recordings circulated by victims on Reddit and X (formerly Twitter) demonstrate that attackers were able to command the support chatbot to associate a new, unauthorised email address with a target account. Once the AI accepted this change, the platform sent verification codes to the attacker’s email, effectively granting them full access. By bypassing traditional identity verification, the hackers were able to reset passwords and lock out the legitimate owners, who were then left stranded in a loop of automated, unhelpful support responses.

High-profile targets and the scale of the breach

The vulnerability did not discriminate between everyday users and global icons. Among the most notable incidents was the temporary hijacking of the official Instagram account associated with former U.S. President Barack Obama’s White House tenure (@obamawhitehouse). The incident served as a stark reminder of how even dormant, high-profile accounts are not immune to automated exploits. As reports of these hacks mounted, it became clear that the issue was systemic, with many users claiming they were not only locked out but subsequently penalised by the platform for the very activity the hackers had initiated.

Meta’s guarded response

Meta’s handling of the situation has drawn significant criticism for a perceived lack of transparency. The company did not issue a formal press release or a detailed incident report regarding the security lapse. Instead, information trickled out through Andy Stone, a Meta spokesperson, who engaged with affected users on the X platform. Stone stated that the company had resolved the issue and was in the process of securing impacted accounts, yet he offered no confirmation on whether the AI support chatbot was indeed the primary point of failure.

The broader implications for platform security

This incident highlights the inherent risks of offloading critical security functions to automated systems. While Meta intended to improve user experience through an AI-driven support interface, the failure to implement robust guardrails allowed bad actors to weaponise the bot’s capabilities. Although reports indicate that Meta has since deployed emergency patches to plug the vulnerability, the event has left many users questioning the safety of their digital identity. As platforms increasingly rely on automation, the trade-off between user convenience and security remains a precarious balancing act.