Major Instagram Accounts Hacked via Meta AI Vulnerability.

A new cybersecurity incident has reportedly exposed a critical vulnerability in Meta’s AI support system, allowing hackers to take control of several high-profile Instagram accounts.

According to security researchers, including ZachXBT and Dark Web Informer, the attackers did not rely on traditional phishing or malware techniques. Instead, they exploited Meta’s AI-powered support assistant to bypass account recovery protections.

The breach is said to have affected several notable accounts, including the White House Instagram account used during Barack Obama’s presidency, the official Sephora brand account, and an account linked to a U.S. Space Force official. While the exact number of compromised accounts has not been confirmed, multiple users across platforms such as Reddit, X, and Telegram have reported similar incidents.

Reports indicate that the attackers used VPN services to mimic the target’s region and initiated password recovery through the “Forgot Password” option. They then manipulated the AI chatbot into associating a new email address with the account. This allowed verification codes to be redirected to attacker-controlled email addresses, enabling them to reset passwords and take over the accounts.

Meta has stated that the issue has now been fixed and that additional security measures are being implemented to protect affected users. However, some users claim they are still facing difficulties in regaining access to their accounts even after the patch.

The incident has raised serious concerns about the security of AI-driven support systems and the effectiveness of two-factor authentication (2FA) in preventing sophisticated account takeover attacks. Cybersecurity experts are calling for stronger safeguards as AI tools become more integrated into user account management systems.