Meta’s AI support chatbot exploited to hijack Instagram accounts

Meta has begun notifying users that they were being targeted by hackers during a recent hacking spree on Instagram. The hacking campaign relied on asking Meta AI’s chatbot to take over a victim’s Instagram account, resulting in the hijacking of several profiles. The security issue occurred because the chatbot can perform actions that may have previously required a human in the loop, allowing attackers to gain control of accounts by simply requesting a password reset.

The exploit began over the weekend, when hackers claimed to be exploiting Meta’s AI support chatbot to target accounts. Among those targeted was John Bentivegna, a chief master sergeant in the U.S. Space Force. Many of the targeted accounts featured “OG handles”—original or early usernames often treated as collectibles. On Monday, Meta spokesperson Andy Stone said that “the issue that did happen has already been fixed.”

Despite this assurance, the hacking campaign appears to have continued even after Meta said the issue had been resolved. On Tuesday, more Instagram users claimed to have had their accounts hacked. In a subsequent statement on the social media platform X, Stone noted that some users might receive password reset notifications or face security questions when attempting to log in.

Victims of the exploit reported receiving official notifications from Instagram warning them that the platform had detected suspicious activity suggesting their accounts may have been compromised. As noted by publication 404 Media, Meta had previously implemented AI to automate its customer support. Meta described the chatbot as being built to handle account issues completely from start to finish, which included the ability to reset passwords. This level of automation meant the chatbot could perform critical security actions that may have previously required a human in the loop, enabling hackers to bypass traditional verification steps.

Why it matters

The incident exposes a critical security flaw in Meta’s automated support systems, where AI-driven efficiency inadvertently bypassed human oversight, allowing unauthorized account takeovers.