The Pulse

Vals AI joins expanded NIST AI Consortium

Vals AI is participating in the reconvened NIST AI Consortium, which focuses on AI measurement, innovation and adoption. The company already provides benchmarks and associated data for NIST AI model evaluations.

AI.info Team · · fact-checked · against 4 sources ↓ · Published 12 Nov 2026

Vals AI is participating in the reconvened NIST AI Consortium, joining a federal research partnership focused on measuring artificial intelligence. The company said in a Nov. 12 post that it would work with the National Institute of Standards and Technology and other members on evaluations. The agency's member list also includes Vals.

NIST broadened the consortium's scope and renamed it in May, changing its name from the AI Safety Institute Consortium. The reorganized consortium is to focus on AI measurement, innovation and adoption through six task groups.

A wider measurement agenda

NIST initially announced the establishment of the earlier consortium in 2023, bringing together more than 280 organizations to develop guidelines and standards for AI measurement. Its expanded work includes building an evaluation ecosystem, investing in AI-enabled science and promoting U.S.-developed AI technology and systems.

One task group is to develop tools for determining whether an AI system meets its design requirements and is suitable for its intended use. The AI Evaluation and Measurement Methods group is to identify and prioritize gaps and open questions in evaluation science, drawing on input from different sectors and types of organization.

Another group will support a toolkit for assessing AI risks and impacts for use with NIST's ARIA program. The AI Documentation Cards group is to develop templates for documenting datasets, models, systems and testing.

A group focused on generative AI limitations will work with IARPA to explore approaches to misinformation, sensitive information leakage, inaccurate reasoning and susceptibility to attack. NIST is also restarting a task group on chemical and biological security to share insights on relevant measurement and evaluation approaches.

Vals' work with NIST

The company is already providing benchmarks and associated data to NIST for AI system evaluation. Those materials are for use by the agency's Center for AI Standards and Innovation in evaluations of AI models before and after deployment.

The company works with domain experts to build benchmarks based on real-world tasks, according to its announcement. It also researches cybersecurity, child safety, and AI's environmental and economic impacts.

NIST does not evaluate commercial products through the consortium and does not endorse products or services used in it, the announcement said.

Participation in the consortium

Co-founder and CEO Rayan Krishnan said Vals looked forward to working with the agency and fellow members on reproducible evaluations that reflect real-world AI use. Glenn Parham, the company's head of public sector, said independent evaluations give policymakers a basis for comparing performance and identifying potential uses.

When it expanded the consortium, NIST invited organizations with relevant technical capabilities to submit letters of interest. Existing members do not need to reapply, but must sign an amendment accepting the changes to the consortium. Member organizations enter into a cooperative research and development agreement with the agency.