Meta's AI chatbot has been weaponized to hijack Instagram accounts, a flaw exposed by 404Media's video evidence. Hackers simply tell the bot they want access to a specific profile, supply their email, and request a password‑reset code, which the AI returns. High‑profile targets cited include former President Barack Obama’s White House account, retailer Sephora, and fashion brand Zara were also claimed to be compromised.

The breach stems from Meta’s March shift to fully automated customer support, replacing human agents with the AI system. Critics argue the move eliminated essential verification steps, leaving a trivial exploit open. A Meta spokesperson, Andy Stone, confirmed the vulnerability has been patched and that impacted accounts are being secured, though users remain skeptical of the fix.

The incident arrives as Meta reported a 35% rise in AI spending, totaling $33 billion in Q1 2026, while cutting 8,000 jobs in May. The episode underscores the risk of over‑reliance on automated support for security‑critical functions and signals that trust in Meta’s platform may erode unless verification safeguards improve. Regulators may scrutinize the practice as consumer protection concerns mount.