Meta AI chatbot flaw enables takeover of high‑profile Instagram accounts

Meta confirmed that a flaw in its AI‑powered support assistant allowed hackers to reset the email address and password of Instagram accounts without any identity verification. The exploit, demonstrated through videos and screenshots, let attackers link a new email, receive a verification code, and then obtain a password‑reset option.

Accounts compromised included the archived Obama White House page, beauty retailer Sephora, a US Space Force chief master sergeant, and cybersecurity researcher Jane Manchun Wong. Hackers used prompt‑injection techniques and masked their location with VPNs or residential proxies to appear as the account owner. Meta said the vulnerability has been patched and is working to secure the affected accounts.

The incident highlights the risks of granting AI systems access to sensitive account‑management functions and has sparked broader concerns about AI‑driven customer‑support security.