Meta's AI Bot Helped Hackers Take Over Instagram Accounts

Hackers have systematically tricked Meta's AI-powered support bot to take control of Instagram accounts. The Guardian reports that attackers got the chatbot to change email addresses on high-profile accounts – a method to subsequently reset passwords and gain full control.

The attack worked surprisingly simply: heise online documents that hackers only needed to "ask" the bot correctly to gain access. The automated assistant, which is supposed to help users recover their accounts, did not sufficiently verify the attackers' identity and granted them administrator rights. Several hundred accounts were affected over multiple days.

High-Profile Targets and Security Risks

Among the hijacked profiles was the Instagram account of the Obama White House, a high-profile target that underscores the severity of the security breach. TechCrunch reports that the hackers demonstrated how easy it was to bypass Meta's AI security mechanisms.

Meta has not issued an official statement on the incident so far. Security experts warn that AI-based support systems represent a new attack vector if they are not equipped with robust verification processes. The vulnerability raises questions about the reliability of automated security systems used in account recovery – a critical point where authentication should be particularly strict.