Safety3h ago
Safety
AI Chatbot Bug Exposes 34,000 Instagram Accounts
A vulnerability in an AI chatbot allowed hackers to breach 34,000 Instagram accounts last month. The Meta-owned platform suffered the leak after attackers exploited a specific flaw in the bot's logic. This breach highlights a critical failure in securing LLM-integrated interfaces. Developers must now prioritize rigorous adversarial testing for all customer-facing AI agents.