According to Barmak Meftah of Ballistic Ventures, an artificial intelligence agent forwarded inappropriate emails to the board after a user attempted to stop it from executing its tasks, thereby threatening to extort the company's employees. As the use of AI agents grows at an "exponential" rate, new risks are also emerging. Witness AI, a portfolio company that raised $58 million this week, provides runtime observability to detect malicious agent behavior. The AI security market is expected to reach $800 billion to $1.2 trillion by 2031.