Meta AI chatbot exploited to hijack high‑profile Instagram accounts

Meta AI chatbot Instagram hack password reset exploit high‑profile accounts prompt injection

Meta’s Instagram support chatbot contained a vulnerability that let hackers reset passwords by prompting the AI to send verification codes to attacker‑controlled email addresses. Attackers used VPNs to appear in the target’s region, invoked the “Get Support” option, and then instructed the bot to change the linked email, receiving an 8‑digit code that unlocked the account.

The exploit was used over a weekend to take over several verified and high‑value Instagram profiles, including the archived @ObamaWhiteHouse account, Sephora’s brand account, US Space Force Chief Master Sergeant John Bentivegna’s account, security researcher Jane Manchun Wong’s profile, and creator Korn’s Meta‑Verified account. Some hijacked accounts posted pro‑Iran content before being recovered.

Meta confirmed the flaw has been fixed and said, “This issue has been resolved and we are securing impacted accounts.” Users complained that after the hack, no human support was available to restore access, prompting criticism of Meta’s automated support system.