NIST’s AI Risk Management Framework provides voluntary guidance for organizations designing, developing, deploying or evaluating artificial intelligence. Released in January 2023, AI RMF 1.0 is intended to help incorporate trustworthiness considerations into decisions about AI systems. It is a risk-management resource, not a certificate that a particular product is safe. NIST framework
The framework organizes its core around four functions: Govern, Map, Measure and Manage. These functions give teams a common structure for discussing organizational responsibilities, the context in which a system operates, how risks are assessed and how those risks are addressed. NIST also published a profile addressing generative AI in July 2024, extending the framework’s application to that class of technology.
Accountability is a related concern. The OECD’s AI principles call for accountability appropriate to each actor’s role and context, including traceability of relevant datasets, processes and decisions across the system’s lifecycle. That makes documentation and responsibility part of governance, rather than treating governance as a one-time technical test. OECD accountability principle
These are published frameworks for evaluating AI practice. They do not establish that any named consultancy has implemented them or independently verified a particular system.
Archive note: the earlier claim of a practice expansion was not substantiated by the cited sources.