NIST
The National Institute of Standards and Technology (NIST) is a federal agency within the United States Department of Commerce, founded in 1901. Its core functions include developing industrial and measurement standards as well as technical guidelines and frameworks for fields such as cybersecurity and artificial intelligence, with the aim of promoting technological innovation and enhancing industrial competitiveness. NIST standards and frameworks are widely adopted by government agencies, industry, and academic research institutions.
Core Facts
The US NIST has published an AI Risk Management Framework that is being adopted by an increasing number of enterprises
The EU Artificial Intelligence Act (AI Act) explicitly requires that high-risk AI systems must have human oversight and intervention capabilities.
NIST发布了AI风险管理框架（AI RMF）
Timeline (last 90 days)
NVD is an enriched database built by the US NIST based on CVE identifiers, adding structured fields such as CVSS severity scores, CPE matching, and affected version ranges.
ML-KEM (formerly known as Kyber) is one of the first batch of post-quantum cryptographic algorithms formally standardized by NIST in the United States in 2024 (FIPS 203), constructed on the Learning With Errors problem over module lattices, and is immune to Shor's algorithm.
NIST defines separation of duties as: 'no single user should be given sufficient privileges to abuse the system alone,' and gives the two-person rule as an example.
P-192, P-224, P-256, P-384 and P-521 were published in 2000 with FIPS 186-2.
An overview of synthetic content risk controls published by NIST notes that a single detector or watermark is insufficient to reliably prove audio authenticity, that watermarks may fail due to compression, resampling or malicious attacks, and that metadata can be stripped or forged.
The AI Standards and Innovation Center under NIST defines agent hijacking as a form of indirect prompt injection.
When each attack was repeated 25 times, the average success rate across five tasks rose from 57% to 80%.
NIST's testing of the Agent Dojo benchmark showed that the strongest attack at the time had an 11% success rate against the upgraded model; NIST and UK AISI then designed new attacks against that model, raising the success rate to 81%.
NIST AI 600-1 defines confabulation as the phenomenon in which a model confidently generates incorrect or false content.
Filippo Valsorda reports that the bounty on the origin of the NIST elliptic curve seeds has been claimed.
40 more timeline events
All Facts (20)
The US NIST has published an AI Risk Management Framework that is being adopted by an increasing number of enterprises
90%VerifiedThe EU Artificial Intelligence Act (AI Act) explicitly requires that high-risk AI systems must have human oversight and intervention capabilities.
90%VerifiedNIST发布了AI风险管理框架（AI RMF）
90%VerifiedThe practical framework for zero-trust architecture is defined by the NIST SP 800-207 standard, with core components including the Policy Engine, Policy Administrator, and Policy Enforcement Point.
80%VerifiedThe U.S. government's risk classification of AI models primarily references the NIST AI Risk Management Framework and the dual-track evaluation mechanism established by the Biden administration's 2023 AI Executive Order.
80%VerifiedThe AI Risk Management Framework published by the U.S. NIST lists transparency as a core element
80%VerifiedAfter a vulnerability enters the NVD (National Vulnerability Database), it is assigned a CVSS score (0–10), where the Critical level corresponds to 9.0–10.0.
75%VerifiedNIST SP 800-207 provides a standard reference for zero trust architecture for federal agencies.
75%VerifiedThe NIST AI RMF was officially released in January 2023, providing a four-dimensional assessment structure of Govern–Map–Measure–Manage.
75%VerifiedThe RBAC permission model was formally proposed by the National Institute of Standards and Technology (NIST) in 1992.
75%VerifiedNVD (National Vulnerability Database) is the U.S. national vulnerability database, where more detailed technical information and affected software versions can be queried.
65%VerifiedThe NIST SP 800-207 standard defines in detail the reference model for Zero Trust Architecture
65%VerifiedSHA-256 is a cryptographic hash function in the SHA-2 family with an output length of 256 bits, possessing collision resistance and preimage resistance.
65%VerifiedGoogle's BeyondCorp and the NIST SP 800-207 Zero Trust Architecture standard provide reference frameworks, but zero trust practices for non-human identities (NHI) are still in early stages of exploration.
65%VerifiedThe LWE-based cryptographic scheme Kyber/ML-KEM has been selected by NIST as the post-quantum key encapsulation standard.
65%VerifiedNIST officially released the post-quantum cryptography standards CRYSTALS-Kyber and CRYSTALS-Dilithium algorithms in 2024.
65%VerifiedThe NIST post-quantum standard suite includes ML-DSA (FIPS 204, derived from CRYSTALS-Dilithium) and SLH-DSA (FIPS 205, derived from SPHINCS+) for digital signatures.
65%VerifiedNIST's post-quantum cryptography standardization project, launched in 2016, officially published its first set of standards in 2024: ML-KEM, ML-DSA, and SLH-DSA.
70%VerifiedNIST formally published three post-quantum cryptography Federal Information Processing Standards (FIPS) in August 2024.
65%UnverifiedRBAC（Role-Based Access Control）是NIST于1992年正式提出的访问控制模型
90%