Hackers Steal Instagram Accounts Using AI-Generated Selfies to Bypass Meta Verification

Attackers have hijacked multiple high-value Instagram accounts by exploiting Meta's AI-powered support assistant, tricking it into transferring ownership using deepfake selfie videos. Among the affected accounts were one previously managed by the Obama White House team, researcher Jane Manchun Wong's @wong account, the @hey handle, and @korn. The account holders report being locked out permanently because Meta's recovery process is entirely automated, offering no path to human support.

The attack chain is technically straightforward. The threat actor initiates the "forgot password" flow, then submits a still photo pulled from the target's public profile, runs it through an AI video generator to produce a short animated clip, and uploads the result as the selfie verification. Meta's AI assistant accepts the synthetic video because, according to victim André, "it can't tell the difference between a real selfie and an AI-generated video of someone's face." Reports also indicate that attackers used VPN services to spoof the account holder's usual region, bypassing geolocation checks that would otherwise trigger a more complex login flow. Once verified, the attacker changes the associated email address, triggers a password reset, and walks away with full control, including 2FA bypass.

The bigger problem is the recovery process. Users describe being trapped in a chatbot loop, with one victim spending six hours receiving nothing but four broken support links. "We're at the point where one AI stole it, and another can't fix it, zero humans in the loop anywhere," the @korn account owner stated. Affected users, including rare single-letter handles like @e and @f, are unable to escalate to a human agent, raising serious questions about Meta's account recovery design. Users concerned about similar risks should audit their exposure by running an email breach checker, testing whether their connection leaks identifying metadata via a VPN/proxy detector, and reviewing their credentials with a password checker to ensure no reused or compromised secrets are tied to high-value social accounts.

The incident highlights a growing class of identity attacks where generative AI defeats biometric verification and where fully automated support systems leave victims with no remediation path. Until Meta reintroduces human-in-the-loop escalation, holders of high-value or rare social media handles should treat their accounts as critical assets: enable hardware security keys rather than SMS-based 2FA, minimize public facial photos, monitor for unauthorized sessions, and be aware that platform AI agents can be socially engineered just like human support staff.