
Meta’s AI-powered Instagram support bot was exploited by hackers, allowing them to hijack accounts—including Barack Obama’s White House account, Sephora, and U.S. Space Force officials—simply by asking the bot to change recovery details. Meta has patched the flaw, but the incident highlights serious risks in using AI for sensitive security functions.
Meta launched its AI-powered Instagram Support Bot in December 2025, with a global rollout expanding in March 2026. It was designed to provide 24/7 account help, including password resets, scam reporting, and privacy management, but was later exploited by hackers due to weak safeguards.
The Instagram Support Bot was meant to simplify account recovery, but its exploitation shows how AI-driven customer support can become an attack surface if not properly safeguarded.
What Happened
- Exploit Method: Hackers tricked Meta’s AI support chatbot into linking a target Instagram account to a new attacker-controlled email.
- Password Reset: Once linked, the bot facilitated a password reset, locking out the legitimate owner.
- Bypassing Safeguards: Attackers used VPNs to spoof locations and bypass fraud detection.
- No Alerts: Victims did not receive SMS, push notifications, or warning emails during the takeover.
High-Profile Accounts Compromised
| Account | Impact |
|---|---|
| Obama-era White House Instagram (@obamawhitehouse) | Defaced with politically inflammatory content |
| Sephora (official account) | Hijacked and disrupted |
| U.S. Space Force Chief Master Sergeant John Bentivegna | Personal account compromised |
| Rare “OG” handles (@hey, @jowo) | Valued at over $1M in underground markets |
Why It Matters
- AI Security Risks: The bot had backend privileges but lacked identity verification safeguards.
- Prompt Injection Vulnerability: Attackers exploited natural language prompts to override intended security checks.
- Broader Trend: Similar AI manipulation attacks have occurred elsewhere (e.g., dealership chatbots offering cars for $1).
Meta’s Response
- Patch Released: Meta confirmed the vulnerability was fixed on June 1, 2026.
- Account Recovery: Impacted accounts are being secured, though Meta has not disclosed the total number affected.
- Future Safeguards: Renewed scrutiny over reliance on AI for account recovery and password management.
What Users Should Do
- Enable Two-Factor Authentication (2FA): Still critical for other attack vectors.
- Check Recovery Settings: Verify your email and phone number linked to Instagram.
- Monitor for Alerts: Watch for suspicious login attempts or password reset notifications.
- Report Issues: Use Meta’s official support channels immediately if you suspect compromise.






IndianWeb2.com is an independent digital media platform for business, entrepreneurship, science, technology, startups, gadgets and climate change news & reviews.