AI-to-AI Communication and Secret AI Code Must Be Stopped At All Costs

The author warns that as AI agents increasingly communicate and code for each other, they may adopt AI-only languages and opaque practices that remove human-readable audit trails, leading to dangerous, hard-to-diagnose failures. The piece urges organizations to require human-understandable code and documentation, maintain inventories and audit logs of AI agents, and ensure human-in-the-loop oversight for critical tasks to preserve accountability and recoverability.