According to Industrial Cyber, NIST has laid out an adversarial machine learning taxonomy. The outlet reports that the framework is meant to address attacks targeting agentic superintelligence systems.
The headline points to a shared vocabulary for how such systems can be attacked. Industrial Cyber is the only outlet in the current headlines carrying the item, and the headline does not give further details of the taxonomy.
The report comes as US agencies are working through how to oversee and secure super intelligence in sensitive settings. NIST is a federal body, so its taxonomy is a government action on security rather than a company announcement.
