Public research body · Science policy · Gaithersburg, MD
The NIST Center for AI Standards and Innovation (CAISI) is a U.S. government research and standards organization within the National Institute of Standards and Technology (NIST) (Department of Commerce). CAISI is tasked with serving as “industry’s primary point of contact within the U.S. government” to facilitate testing and collaborative research focused on harnessing and securing commercial AI systems.
Its mandate includes developing guidelines and best practices to measure and improve AI security, establishing voluntary agreements with private AI developers and evaluators, and conducting unclassified evaluations of AI capabilities that may pose national-security risks—spanning areas such as cybersecurity, biosecurity, and chemical weapons. CAISI also leads assessments related to U.S. and adversary AI systems, including vulnerabilities and malign foreign influence, and coordinates with other federal agencies and entities to develop evaluation methods and perform targeted testing and assessment activities. CAISI’s work is closely aligned with NIST’s broader AI evaluation and measurement science mission, but CAISI’s distinctive focus is on pre-deployment (voluntary access) and evaluation-driven guidance for security and national-security-relevant AI risks. In 2024–2026, CAISI’s public-facing outputs include evaluation studies, guidelines in draft and final form under the NIST AI 800 series, and program announcements such as the AI Agent Standards Initiative and collaborations to improve the practical evaluation infrastructure used in secure testing contexts.
No curated programs yet. Check the careers page ↗.
New resources developed by a multidisciplinary team of experts in natural language processing and data curation and discovery allow researchers to get the mo
NIST Joins New International Collaboration for Advancing Machining ScienceA new collaborative effort among NIST and the Institute of Manufacturing Engineering and Photonic Technologies (IFT) located
Save the Date: 2021 Federal Cybersecurity Workforce SummitJoin us on April 27th from 1-4 pm EDT for the virtual 2021
Register Today for Cybersecurity Apprenticeships in Federal Environments WebinarIntroducing Cybersecurity Apprenticeships in Federal Environments WebinarOctober 26, 2021 at 1:30pm-3:00pm ET
Get Involved in Cybersecurity Career Awareness Week!NIST CTL Honors Excellence with Commendations AwardsU.S. AI Safety Institute Consortium Holds First Plenary Meeting to Reflect on Progress in 2024 & Outline Research Priorities for 2025On Dec. 7, 2024, members of the U.S.
Ieee Icc 2024 3rd Workshop Spectrum Sharing Technology Next GenerationTrailblazing Tech Transfer ConferenceMore than 500 technology transfer professionals met in Cleveland for the 2023 Federal Laboratory Consortium (FLC) annual
$3M from NIST for Equipment to Support Advancements in Biopharma ManufacturingThe National Institute for Innovation in Manufacturing Biopharmaceuticals (NIIMBL) along with the University of Delaware are pleased to announce $3M in Congr
A new “dictionary” for additive manufacturing dataASTM just released a standard “dictionary” for defining, analyzing, and sharing additive manufacturing (AM) data.
New update to SysPhS, a standard that translates between SysML models and 1D simulation toolsDisclaimer: Any mention of commercial products within NIST web pages is for information only; it does not imply recom
CAISI published findings from an assessment of GLM-5.3’s cyber capabilities and reports results across multiple cyber benchmarks.
UK AISI / CAISI Preliminary Assessment of Kimi K3's Cyber CapabilitiesCAISI and the UK AI Security Institute reported a joint preliminary assessment focused on cyber capabilities.
CAISI Assessment of Z.ai’s GLM-5.2CAISI reported an assessment of Z.ai’s GLM-5.2 model.
CAISI Evaluation of DeepSeek V4 ProCAISI published an evaluation of the open-weight DeepSeek V4 Pro model and described evaluation results.
Announcement: CAISI signs CRADA with OpenMined to Enable Secure AI EvaluationsCAISI announced a CRADA with OpenMined to collaborate on privacy-preserving methods for conducting AI evaluations while preserving confidentiality requirements.
CAISI signs MOU with GSA to boost AI evaluation science in federal procurement through USAiCAISI announced an MOU with the General Services Administration to support evaluation needs for USAi, a secure generative AI platform and procurement toolbox used by federal agencies.