NIST Is Seeking Input on Federal Standards on the Use of Artificial Intelligence
Privacy in Focus®
On May 1, the National Institute of Standards and Technology (NIST) within the U.S. Department of Commerce published a request for information (RFI) on an anticipated plan for federal engagement on standards for artificial intelligence (AI). The agency is seeking input on topics that include the technical standards and guidance that are needed to establish “trustworthy” aspects of AI technologies, including “accuracy, transparency, security, privacy, and robustness.” The agency will also be hosting a workshop on May 30 to obtain public input on AI standards.
NIST’s development of a federal plan on standards and tools to support “reliable, robust and trustworthy” AI systems is a central component of the Administration’s Executive Order on AI, and will provide an important framework for the federal approach to AI going forward. We discuss the Executive Order in greater detail here. The workshop and comment period provide a key opportunity for stakeholders to engage on how the U.S. government defines and begins to address potential AI regulatory issues, which include how data is used and managed, whether AI processes need to be “explainable,” and how to address potential bias concerns. NIST’s work will be influential both domestically and in setting a course on international engagement on AI standards.
Comments in response to the RFI must be filed by May 31, 2019, and in-person registration for the workshop closes on May 23.
The RFI
NIST seeks information to help develop “[t]imely and fit-for-for purpose AI technical standards,” which NIST believes will be crucial for building trust in AI technologies and achieving economies of scale. In general, NIST seeks to understand:
- The current status and plans regarding the availability, use, and development of AI technical standards and tools;
- Needs and challenges regarding the existence, availability, use, and development of AI standards and tools; and
- The current and potential future role of federal agencies regarding the development and use of AI technical standards and tools.
The RFI also provides a list of topics covering the “major areas” of inquiry, which include:
- The need for AI technical standards and the challenges in identifying and developing them;
- Whether AI standards should be sector-specific or there should be cross-sector AI standards;
- Technical standards and guidance that are needed to establish “trustworthy” aspects of AI technologies, including “accuracy, transparency, security, privacy, and robustness;”
- Opportunities for and challenges to U.S. leadership in AI standards development;
- Federal agencies’ needs for AI standards and involvement and their current involvement in standards development; and
- Actions the federal government can take to help ensure that desired AI standards are incorporated into practice.
The Workshop
NIST has also scheduled a workshop for May 30 to discuss its plan for federal engagement on AI standards. The workshop will include both panels and working sessions. Topics will include the importance of effective U.S. leadership on AI standards, what technical standards already exist, the needs of federal agencies, and how the federal government can best engage on AI standard-setting.
NIST expects to release an annotated outline after the workshop concludes. Under the Executive Order, the plan must be finalized by August 10 of this year.
NIST’s engagement plan will be one of the first substantive moves by the federal government to tackle potential regulatory issues surrounding AI. Much as with its approach to cybersecurity, NIST’s approach to AI standards could significantly influence how these technologies are adopted and used. The workshop and RFI are important opportunities for stakeholders to engage prior to the August 10 deadline for the final plan.
© 2019 Wiley Rein LLP
