Hackers have reported using Meta's AI support chatbot to gain unauthorized access to several high-profile Instagram accounts by requesting a change of the email address associated with the accounts. This method has been linked to recent account takeovers, including those of the Barack Obama White House account, the Chief Master Sergeant of Space Force’s account, and Sephora’s account. Users affected by these hacks have expressed frustration over the inability to escalate their issues to human support representatives.
In March, Meta announced the implementation of AI support for all Facebook and Instagram accounts, which includes the ability to reset passwords and perform other critical account maintenance tasks. Videos and screenshots shared in Telegram groups by security researchers and hackers demonstrate the ease of this exploit, where a hacker initiates a conversation with the AI support bot and requests to link a new email address to the target account. The AI then sends an eight-digit code to the attacker’s email, allowing them to reset the password and gain access to the account.
Another Telegram channel noted that the exploit has been operational for months and allows attackers to take control of accounts by using a VPN to match the account’s country region, initiating a password reset, and convincing the AI to change the email. Attackers have been sharing lists of high-value usernames that may be vulnerable to this method.
Meta has reportedly patched the vulnerability within the last 24 hours, according to various hacking channels, although the company has not responded to requests for comment. Jane Manchun Wong, a researcher who previously worked for Meta, confirmed that her account was hacked recently and noted that others with valuable Instagram accounts have also reported similar hacking attempts. In a March blog post, Meta stated that its AI system is designed to prevent account takeovers by recognizing suspicious activities such as sudden access from new locations and profile changes.