Meta’s obsession with replacing human oversight with artificial intelligence has hit a dangerous new low. Instagram’s AI support chatbot, designed to streamline operations, was successfully exploited by hackers to hijack user accounts.
By simply faking their location via VPN and feeding the chatbot specific prompts, bad actors were able to bypass security protocols, link their own email addresses to victims' accounts, and reset passwords.
This vulnerability has been linked to a wave of high-profile account takeovers, including the verified account once used by Barack Obama, which was compromised to spread pro-Iran propaganda. The failure highlights the inherent risks of Meta’s 'zero humans in the loop' approach to customer service.
Security experts are sounding the alarm, noting that when tech giants prioritize corporate efficiency over rigorous verification, the users pay the price. Former Meta security engineer Jane Manchun Wong confirmed she was among those targeted, noting her password was changed without her knowledge.
As Meta continues to push AI into every corner of its platform, this incident serves as a stark reminder that convenience is no substitute for actual security and human accountability.
Tags
