Artificial agents can and should whistleblow, but only within a normative framework rooted in human whistleblowing traditions. The paper calls for government regulators to establish clear guidelines on what machines may disclose and how to legally protect developers of such systems.