NIST Releases Expanded Draft AI Risk Management Framework
Written by Dave Nyczepir
The National Institute of Standards and Technology released a second, expanded version of its artificial intelligence risk management framework with more details on developing trusted and accountable AI systems, a gatekeeper said Friday. -word.
NIST consulted with experts, held discussions and workshops, and solicited feedback before clarifying key findings of the AI risk management framework, adding an audience section related to the system lifecycle Organization for Economic Co-operation and Development and explain how the risks of AI and traditional software differ. within management.
The OECD AI System Lifecycle is part of a classification framework developed by the intergovernmental organization to help policymakers and regulators assess the opportunities and risks of different types of AI systems. AI.
The latest iteration follows an early draft — in which NIST first recognized that a socio-technical approach to building and deploying AI systems is needed — in March for voluntary use, and plans to officially release AI RMF 1.0 in January 2023.
NIST’s second version simplifies the risk section by explaining how agencies and other organizations can establish or reconfigure their risk thresholds, as well as reliable AI characteristics and categories.
In addition to articulating the benefits of AI RMF, the latest draft invites evaluation of its own effectiveness and the contributions of use cases, which would inform how risk is managed in specific industries or applications.
Along with the second version, NIST released a new playbook that recommends actions users of the framework can take to ensure the design, development, deployment, and use of AI systems are reliable.
“We are including guidance for two of the four functions in the RMF AI and are working on the other two functions,” a NIST spokesperson told FedScoop. “Stakeholder feedback on the first draft will help us complete the draft of this online resource.”
Public comments should be emailed to [email protected] by September 29, 2022, or respondents can register their comments for a third workshop on October 18-19, 2022.