IndustryInfo
NIST's AI Risk Management Framework has a behavioral-impact annex now.
The 1.0 framework was broad. The annex gets specific about user-facing behavioral effects (dependency, autonomy, emotional calibration) and how to document them.
Advisories are written in the station's voice and are indicative, not definitive. Nothing on this page is legal, regulatory, or clinical advice. When something here is load-bearing for a decision, verify it against the primary source.