Voluntary framework from the United States National Institute of Standards and Technology for managing AI risks. It organises the work into four functions: govern, map, measure and manage.