Concerns surrounding the security of Meta's AI support chatbot have intensified following reports of high-profile account breaches, including those of Barack Obama and beauty retailer Sephora. The incidents were highlighted by 404 Media, revealing that hackers exploited the AI assistant to gain unauthorized access to Instagram accounts.
Meta has acknowledged the issue, stating that the problem has been addressed and that they are taking steps to secure affected accounts. The company did not disclose how many accounts were compromised, but users reported similar hijackings on platforms like Reddit and X. Videos and instructions on how to bypass security measures were circulated on Telegram, showcasing the process of linking accounts to new email addresses through the AI assistant.
Earlier this year, Meta introduced the AI support assistant across its platforms, aiming to enhance user support by allowing actions such as password resets and reporting scams. This initiative is part of a broader strategy under CEO Mark Zuckerberg, who has committed $145 billion to AI development, including infrastructure investments. The recent breach raises significant questions about the reliability of AI in safeguarding sensitive user information.