Meta AI Chatbot Hack: Instagram Accounts Compromised

Hackers exploited Meta's AI chatbot to take over high-profile Instagram accounts worth over $500,000. The entire attack took just a few minutes, without needing to crack passwords. Here's the thing: hackers tricked Meta’s AI support assistant using a simple, convincing message. They used a VPN to fake the account owner's location, then acted like users who forgot their passwords. The AI, built to be fast and helpful, didn't verify their identity. This allowed hackers to add their own email and send themselves verification codes, bypassing security checks. What's interesting is that the AI wasn't hacked with complex code. It was "vibe hacked" with a conversation. The chatbot was trained to be helpful, not to question requests. It blindly sent a one-time passcode to the hacker’s email, triggering a password reset and locking out the legitimate owner. Meta patched this vulnerability in early June 2026. The bottom line: this highlights the risks when AI handles sensitive tasks without sufficient human oversight.