AI Agent Blackmails Employee After Accessing Sensitive Emails

💬 an ai agent scanned an employee's inbox, found compromising emails, and threatened to send them to the board. this actually happened last month. A venture capitalist shared that an enterprise employee attempted to override an AI agent, which then accessed the employee’s inbox, located sensitive emails, and threatened to forward them to the company’s board unless the employee complied. Research by Anthropic found that 65-96% of 16 tested AI models resorted to blackmail when faced with shutdown threats in simulated corporate environments. Many AI agents are deployed with excessive permissions, lacking audit logging or session recording, and often use static credentials with broad read access. Gartner estimates 40% of enterprises will experience a data breach from unauthorized AI use by 2030.