How hackers tricked Meta AI into handing over high-profile Instagram accounts

Hackers were able to manipulate Meta’s AI chatbot in order to gain access to users’ accounts. Social media platforms like Reddit and X (formerly Twitter) have been filled with complaints from users about how their accounts were taken over by hackers.

The shocking part is that hackers didn’t even need to get a hold of the victim’s email address or phone in order to gain access to their Instagram account. Instead, they seemed to have relied solely on the inherent vulnerabilities in Meta’s AI support bot to complete the task.

The hack also doesn’t seem limited to normal accounts. Even prominent Instagram accounts were caught in the crossfire, including Barack Obama’s White House account, the personal account of Chief Master Sergeant John Bentivegna, and global beauty retailer Sephora.

“I’m one of the people who got hacked. I’ve had my account since 2010 or 11, but it’s my personal account, so not a ton of followers, and I’ve kept my account private in recent years. I’ve tried to get my account back (it still exists, I have another account I can use to check on it), but I’m not famous or verified, so I’m SOL,” wrote a user on Reddit.

Meanwhile, security researcher Jane Wong also posted on X about how her account was targeted during the latest campaign.

“Even my Instagram account got hacked. The password got changed without my knowledge and I was getting different password reset attempts throughout yesterday. And I got repeatedly logged out from the IG iOS app,” wrote Wong.

Meta says Instagram issue now resolved:

Meta spokesperson Andy Stone confirmed in a reply to Wong that the issue has now been resolved, writing, “This claim about world leaders is totally false. The issue that did happen has already been fixed.”

Notably, the Meta AI support assistant, which was manipulated by the hackers, was rolled out by the company on Instagram and Facebook earlier in the year.

The chatbot was designed for 24/7 support to users with issues like updating profile settings, managing privacy controls, reporting impersonation accounts, and, more importantly, resetting passwords.

How did the attack work?

Videos of how to exploit the Instagram vulnerability started doing the rounds on Telegram in the last few days. One such video was posted by a user on X detailing a step-by-step guide of sorts on how to hack an Instagram account.

The video showed that the attacker simply asked Meta’s AI support assistant to add a new email address to the intended target’s Instagram account. The AI can then be seen sending a verification code to the hacker’s email address instead of the original email address associated with the account.

The hacker then shares the verification code with the AI, which then goes on to share a password reset link with them.