In a startling new cybersecurity trend, hackers have bypassed traditional security measures by manipulating Meta’s automated AI customer support systems to take over user Instagram accounts.
As social media platforms increasingly rely on artificial intelligence to handle customer service and account recovery, cybercriminals have found a new backdoor. Instead of trying to steal passwords or breach two-factor authentication directly, attackers have been “social engineering” the AI chatbots designed to help users regain access to locked accounts.
By feeding these AI support systems manipulated prompts and synthetic data, malicious actors successfully convinced the automated agents to hand over control of legitimate profiles. This incident highlights a growing, unintended vulnerability in the tech industry. As companies rush to replace human support agents with cost-saving AI, they are inadvertently creating powerful new attack vectors for bad actors who exploit the AI’s programmed helpfulness.
Following the revelation of the vulnerability, the company acknowledged the ongoing breach and intervened to halt the automated hijackings. “This issue has been resolved and we are securing impacted accounts,” said Meta spokesperson Andy Stone in a statement on X.
Security experts warn that despite this immediate fix, the exploit represents a fundamental shift in cybercrime. Users are encouraged to remain vigilant about their account security, while tech giants face mounting pressure to rethink their AI governance and ensure that automated efficiency does not come at the cost of basic user safety.
