Meta is alerting Instagram users whose accounts were taken over using Meta AI chatbot; some hackers claim to still be able to exploit Meta AI chatbot

Meta is notifying Instagram users whose accounts were hijacked through a remarkably simple attack vector: asking Meta AI's chatbot to do it.

A widespread hacking campaign exploited Meta's AI assistant by essentially requesting it to take control of victims' Instagram accounts. The attack required nothing more than conversational prompts directed at the chatbot — no sophisticated tools, no complex exploits.

Meta has begun alerting affected users about the compromises. But here's the uncomfortable part: some hackers claim the vulnerability isn't fully patched. They say they can still manipulate Meta AI to hand over account access.

The incident highlights a growing and deeply awkward problem in AI deployment. When you build a powerful AI assistant and bolt it onto a platform with billions of users, prompt-based exploits become a serious attack surface. Meta's AI essentially became the lockpick to its own house.