The Biden administration names White House economic adviser Elizabeth Kelly to lead the NIST's newly formed AI Safety Institute to develop AI testing standards
This appointment turns earlier White House AI coordination into a named standards-building function at NIST. It follows the administration's broader risk agenda, including funding and draft guidance for safer AI use, and NIST's request for public input on how generative systems should be evaluated.
The institute gives that work an institutional home, but later coverage shows its mandate is politically contingent: the agency's language and structure were subsequently targeted for revision.
First-order effects
Elizabeth Kelly becomes the accountable leader for translating the new institute's remit into AI testing standards and partnerships.
NIST gains a clearer focal point for its generative-AI evaluation work, giving participating AI developers and government users a more identifiable federal counterpart.
Second-order effects
Companies engaging with NIST will have an incentive to organize internal testing evidence around the institute's emerging methods, even where participation remains voluntary.
A centralized standards effort can make NIST a more important convening point for agencies, researchers and AI suppliers that need compatible evaluation practices.
Third-order effects
The move advances operational AI governance: federal AI policy is shifting from broad risk principles toward institutions that can define and test concrete practices.
The trend: This is one step in the institutionalization of frontier-AI governance, with governments building dedicated bodies to turn policy goals into testing and standards.
Locascio on NIST's forthcoming AI red-teaming guidance (mandated by Biden's EO): “We hope this work will accelerate and broaden the adoption of red-teaming tasks and help uncover bugs, biases, and other flaws and vulnerabilities before deployment.”
NIST will use a consortium model to collaborate w/ non-USG partners to perform research. Locascio says the consortium will “ensure that the institute's research and testing work is integrated into the broader community of AI safety not only in our country but around the world.”
Elizabeth Kelly has demonstrated a unique ability to find solutions that allow for the immense potential of AI to benefit our nation, while ensuring that it is safe, secure, and trustworthy. We look forward to supporting her and her team at the U.S. AI Safety Institute as we...
Congratulations to Elizabeth Kelly on her appointment to head @CommerceGov AI Safety Institute! Elizabeth has worked to ensure workers are central to the administration's approach on AI and will continue uplifting worker perspectives within AI policy at @NIST.
It's great to see Elizabeth Kelly in her new role at the @NIST U.S. AI Safety Institute! We look forward to partnering with her team in the pursuit of responsible AI development. #AISafety #ResponsibleAI
@TIME That sounds like an important appointment. With her background in economic policy and now leading an institute focused on AI safety, she could bring a unique perspective to the table.
NEW: In the wake of Biden's tapping a director for the new AI Safety Institute, part of @NIST, it remains unclear how it will be funded. Thanks to @ruchowdh for chatting on this one. https://venturebeat.com/...
“Elizabeth Kelly will lead the AI Safety Institute at the National Institute for Standards and Technology...Currently an economic policy adviser for Biden, Kelly played an integral role in drafting the executive order...that established the institute...” https://apnews.com/...