Meta’s AI chat support within Instagram underwent an exploit over the weekend, leading to the hijacking and password reset of several high-profile accounts. After multiple complaints surfaced on social media, Instagram spokesperson Andy Stone claimed that the issue has been fixed without getting into elaborate details on how the incident transpired.
Meta has been expanding AI-fueled features on its family of apps like WhatsApp and Instagram for over a year now. It essentially exposed over a billion Meta users to AI internationally. This rather aggressive AI push by Meta, however, backfired last weekend when malicious actors targeted its AI customer support systems against the very users that were promised more protection.
Here’s how the attack took shape
Jane Machun Wong, a famed reverse engineer and tech sleuth, was among those who lost access to their Instagram accounts. On X, Wong said, she kept getting back-to-back notifications on password reset attempts. She also added that her Instagram app on iOS repeatedly kept logging her out.
Wong came forward with her own experience of this incident responding to another individual talking about his experience. Andre, who goes by the handle @oracles on X, claimed that he owned a few rare profiles himself and was stressed watching the incident unfold.
“People losing handles they’ve owned since 2010, some worth hundreds of thousands. Obama White House account got hit. These aren’t some random new accounts, these are verified, locked down accounts and they still got compromised,” he said.
Form what it appeared, Andre said, the attackers selected the accounts they wished to target and navigated to their “forgot password” page. There they submitted the reason of a possible account hack, used VPN to match target’s location, and used AI to generate a selfie video of the victim through any of their public pictures to get access to the accounts.
Video representations of how this attack could have processed are also surfacing on social media.
Meta has neither confirmed nor denied the claims even after Wong joined Andre in the discussion.
Meta’s AI > humans decision grabs spotlight
The incident has come to light just weeks after Meta laid-off over 8,000 employees to ramp-up internal AI adoption.
Following the incident, Meta has found itself on the recieving end of rather mocking questions asking if it fired the security team for an AI agent because grievances around AI support not being able to deliver the proper assistance is also getting highlighted.
“Meta’s AI just accepts it (the video) because it can’t tell the difference between a real selfie and an AI-generated video of someone’s face. Once verified they change the email to theirs. Password reset link goes to their email,” Andre’s autopsy of the attack noted. “One AI fooling another AI while there’s literally no person anywhere to catch it.”
The account of the U.S. Space Force’s chief master sergeant John Bentivegna was also reportedly compromised. For now, it remains unclear exactly how many users were impacted by the incident, how many were able to recover their passwords, and what was the complete effect of the breach.
While a detailed post from Meta on the incident remains awaited for now, Instagram’s Andy Stone claimed, “This claim about world leaders is totally false. The issue that did happen has already been fixed.”
