NIST AI Risk Management Framework
An AI risk management framework released by the U.S. National Institute of Standards and Technology. It divides AI risk management into four core functions: Govern, Map, Measure and Manage, and emphasizes continuous risk identification, assessment and treatment.
Timeline (last 90 days)
The U.S. federal government's acquisition of AI tools typically must go through reviews such as FedRAMP authorization, FISMA security requirements and the NIST AI Risk Management Framework.
Red-teaming exercises are gradually being incorporated by regulators and AI companies into pre-release safety evaluation processes for models.
Under the EU AI Act and the U.S. NIST AI Risk Management Framework, a model's ability to refuse harmful instructions is regarded as a positive indicator of high safety standards, whereas the same capability may be characterized as an operational risk under defense procurement logic.
The EU AI Act and the NIST AI Risk Management Framework still have gaps in specific standards for reporting misalignment incidents
All Facts (4)
Red-teaming exercises are gradually being incorporated by regulators and AI companies into pre-release safety evaluation processes for models.
70%UnverifiedThe U.S. federal government's acquisition of AI tools typically must go through reviews such as FedRAMP authorization, FISMA security requirements and the NIST AI Risk Management Framework.
50%UnverifiedUnder the EU AI Act and the U.S. NIST AI Risk Management Framework, a model's ability to refuse harmful instructions is regarded as a positive indicator of high safety standards, whereas the same capability may be characterized as an operational risk under defense procurement logic.
50%UnverifiedThe EU AI Act and the NIST AI Risk Management Framework still have gaps in specific standards for reporting misalignment incidents
50%