Hackers use Meta AI chatbot to hijack Instagram accounts

Hackers have exploited Meta's artificial intelligence chatbot, using it as a tool to seize control of Instagram accounts. The security breach highlights emerging risks as companies increasingly integrate AI into customer support, particularly when chatbots handle sensitive actions like identity verification or account recovery.

International tech outlets reported that attackers manipulated Meta's chatbot to add new email addresses to target accounts. This allowed them to use the standard password reset process to gain full control. Notably, this method did not require malicious software, phishing links, or access to the victim's original email. Instead, hackers leveraged the platform's automated support system, tricking the chatbot into performing actions that should have required stricter identity checks.

Among the affected accounts were prominent ones, including the former White House Instagram account from the Obama administration, Sephora's account, and that of a senior US Space Force official. Some compromised accounts were reportedly altered or used to post irrelevant content. Meta stated the issue has been resolved and affected accounts are being secured. However, the incident prompts questions about whether AI should be entrusted with sensitive steps in account recovery processes.

Security experts describe this as a form of AI manipulation, akin to prompt injection. While typical chatbots might provide incorrect answers, when connected to systems that can alter account information, reset emails, or assist with password recovery, a faulty response can escalate into a significant security incident. This breach underscores the blurring lines between customer support and account security, suggesting that automated AI processes require more robust safeguards than relying solely on a chatbot's contextual understanding.