Our CEO Camellia Chan shared her latest thoughts on AI agent security in Forbes Technology Council, arguing that the industry is falling into a familiar and costly pattern.
Autonomous AI agents have entered production environments with access to databases, API keys, and system privileges — yet the security frameworks built around them relied on the same software-layer approach that failed to prevent breaches across every major technology wave, from network security in the 1990s to cloud infrastructure in the 2010s.
Camellia outlined why agentic AI represented a fundamentally different threat model, where compromised agents could exfiltrate data and move laterally across systems at machine speed — with no human in the loop. She made the case that true defense-in-depth required a hardware layer operating outside the software trust boundary, and that the industry needed to build it before the first major breach, not after.
Read the full article on Forbes Technology Council: https://www.forbes.com/councils/forbestechcouncil/2026/03/20/ai-agent-security-is-repeating-the-industrys-oldest-mistake/
