Last reviewed
Jun 3, 2026
Sources
10 citations
Review status
Source-backed
Revision
v1 · 1,461 words
Add missing citations, update stale details, or suggest a clearer explanation.
The NIST Artificial Intelligence Consortium is a public-private partnership run by the U.S. National Institute of Standards and Technology (NIST) to develop measurement science, evaluation methods, and standards for artificial intelligence. It is the renamed successor to the Artificial Intelligence Safety Institute Consortium (AISIC), which NIST established in 2023. On May 29, 2026, NIST published a Federal Register notice retitling the group, dropping the word "safety," broadening its stated focus toward AI measurement, adoption, and innovation, and reopening membership to new organizations. [1][2]
The rename is the second time in roughly a year that a NIST AI body has shed "safety" from its name. In June 2025 the U.S. AI Safety Institute was reconstituted as the Center for AI Standards and Innovation (CAISI). Both changes are widely read as part of a broader shift in U.S. AI policy under the Trump administration, away from the risk-focused framing of the Biden era and toward promoting American AI development and adoption. [3][4]
NIST announced the formation of the Artificial Intelligence Safety Institute Consortium through a Federal Register notice dated November 2, 2023 (88 FR 75276), shortly after President Biden signed his October 2023 executive order on "safe, secure, and trustworthy" AI. [2][5] The consortium was formally launched on February 8, 2024, when Commerce Secretary Gina Raimondo announced that more than 200 organizations had joined as founding members. The roster included most of the major AI developers, among them OpenAI, Anthropic, Google, Meta, Microsoft, Nvidia, IBM, Hugging Face, Databricks, and Cohere, alongside universities, civil society groups, and other companies. [6]
AISIC was housed under the U.S. AI Safety Institute and tasked with building "science-based and empirically backed guidelines and standards" for AI measurement. Its early work mapped directly onto priorities in Biden's executive order: guidance on red teaming, capability and risk evaluations, AI safety and security, and techniques for watermarking synthetic content. [6] Over its first two years the consortium grew to more than 280 member organizations and produced foundational work toward what NIST describes as global AI metrology, the science of measuring AI system behavior. [1][2]
The May 29, 2026 Federal Register notice (FR Doc 2026-10779) retitled AISIC as the NIST Artificial Intelligence Consortium, revised the scope of its research, and reissued the invitation for organizations to submit letters of interest. [1][7] NIST framed the change as reflecting the group's expanded goals rather than a reduction in activity. In the agency's words, the consortium now concentrates on "AI measurement, innovation and adoption" instead of leading with safety. [1][8]
The stated focus areas include advancing AI measurement science and evaluation, building out an AI evaluation ecosystem, supporting AI-enabled science, and promoting the adoption of AI technology developed in the United States. [1] Deputy NIST Director Craig Burkhardt described the relaunch in capability-focused terms: "We are inviting technically capable organizations to join the NIST AI Consortium to address the challenges associated with the development and deployment of AI." [1][8]
The rebrand follows a directive in the Trump administration's AI Action Plan, released in July 2025, which called for the consortium to convene industry on measurement science that would "promote the development of AI." [3] Reporting by Federal News Network noted that the consortium's relaunch aligned the group's identity with the earlier renaming of the institute that houses it, completing the move away from safety-centered language across NIST's AI portfolio. [8]
NIST organizes the consortium's technical work into six task groups, some carried over from AISIC and at least one restarted for the new phase. [1]
| Task group | Focus |
|---|---|
| AI Testing, Evaluation, Verification and Validation (AI TEVV) | Tools and methods to verify that AI systems meet design requirements |
| Annotation for AI Risks and Validity | Science-based toolkits for assessing AI risks, working with NIST's ARIA program |
| AI Evaluation and Measurement Methods | Identifying gaps and barriers in the science of evaluating AI |
| Bias Effects and Notable Generative AI Limitations (BENGAL) | Reliability of large language models, in partnership with IARPA |
| AI Documentation Cards | Standardized templates for documenting AI systems |
| Chemical and Biological Security | Domain-specific risks from AI applications in chemistry and biology |
Participation works through Cooperative Research and Development Agreements (CRADAs) between each member and NIST. Existing members are asked to sign amendments to reflect the new name and scope, while new applicants enter fresh agreements. [1] The original AISIC membership, drawn from the frontier model developers and the wider AI industry, largely carries forward into the renamed body. [1][6]
The 2026 notice reopened the consortium to new members. NIST said it would accept letters of interest on an ongoing basis, with regular review periods for applicants. The first review period was set to begin within 60 days of the notice's publication, with selection decisions communicated to applicants afterward; subsequent reviews are expected on a roughly biannual schedule. [1][7] Coverage of the relaunch described NIST as accepting applications on a first-come, first-served basis as it works to expand its measurement efforts by drawing on the broader community's capabilities. [7][8]
The consortium's rename is best understood alongside the renaming of its parent institute. The U.S. AI Safety Institute had itself been created under the Biden administration to carry out testing and evaluation of advanced AI models. In June 2025, Commerce Secretary Howard Lutnick announced that the institute would become the Center for AI Standards and Innovation. Lutnick cast the change as a corrective to regulatory overreach, saying that "for far too long, censorship and regulations have been used under the guise of national security," and that the new center would "evaluate and enhance U.S. innovation of these rapidly developing commercial AI systems while ensuring they remain secure to our national security standards." [9][4]
That pattern, replacing "safety" with "standards," "innovation," and "security," runs through the administration's approach to AI governance. Trump rescinded Biden's 2023 AI executive order shortly after taking office in January 2025, and reporting indicates the White House shelved a separate draft order that would have created an FDA-style pre-release evaluation process for AI models. The July 2025 AI Action Plan reframed federal AI efforts around removing what the administration calls barriers to American AI leadership. The consortium rename slots neatly into that agenda. [3][4]
NIST's own AI Risk Management Framework and its broader work on AI standards continue under the new institutional names, so the substantive technical mission, measuring and evaluating AI systems, has not disappeared. What changed is the language and the emphasis. [1][3]
The Trump administration and parts of industry have welcomed the innovation-first reframing as a way to reduce regulatory friction. Civil society groups and several academics have read the same changes as a meaningful narrowing of mandate rather than a cosmetic update.
In a July 2025 analysis written before the consortium's own rename, Paulo Carvão, a senior fellow at the Harvard Kennedy School, together with co-authors Mizuki Yashiro and Shaurya Jeloka, argued that the shift from "safety" to "standards and innovation" was not merely semantic. They contended that "language is never neutral" in AI governance and that the move codified a pivot away from a broad commitment to minimizing harm toward a narrower focus on security and operational concerns. They also noted that a large majority of public comments in the relevant rulemaking had prioritized fairness, accountability, and safety, while industry submissions dominated the resulting policy direction. [10] News coverage of the May 2026 consortium notice itself recorded no organized opposition at the time of publication; the debate has centered on the institute-level rebrand and the underlying policy posture rather than the consortium notice in isolation. [4][10]