Hackers have reportedly exploited Meta’s AI-powered support chatbot to gain access to several high-profile Instagram accounts, including the former White House account from the Obama administration, which still has more than 2.4 million followers. Other targeted accounts include that of the Chief Master Sergeant of the US Space Force and beauty retailer Sephora.
Cybersecurity researchers sharing videos in Telegram groups have demonstrated how the attack may have been carried out. In one clip, a hacker asks the Meta AI assistant to reset a target account’s password, instructing it to send the verification code to a new email address controlled by the attacker. When asked to verify identity with a selfie video, the hacker allegedly presented a fake video generated using AI tools, allowing the email change to be completed.
The method, described by researchers as a form of social engineering, appears to have bypassed two-factor authentication protections. It is not yet known how many accounts were affected, but Meta has acknowledged the issue. In a post on X on Monday, Meta’s head of communications Andy Stone said: “This issue has been resolved, and we are securing impacted accounts.”
The incident raises concerns about social media platforms shifting account support tasks to AI chatbots. In March, Meta introduced AI-powered assistance for all Facebook and Instagram accounts, promoting the tool’s ability to help users reset passwords and perform other account maintenance functions. The feature’s product page stated: “When something goes wrong with your account, you shouldn’t have to hunt for answers.” However, users whose accounts were compromised complained that there appeared to be no way to escalate problems to a human representative.
