CAISI Evaluation of DeepSeek AI Models Finds Shortcomings and Risks
Source: NIST AI · Sarah Henderson
Intel Summary
The U.S. National Institute of Standards and Technology's Center for AI Standards and Innovation (CAISI) has released an official evaluation of multiple artificial intelligence models from China-based DeepSeek. The assessment highlights specific technical shortcomings, safety risks, and performance limitations. This review provides an authoritative U.S. government technical analysis regarding the operational reliability and risk profiles of DeepSeek's widely used model portfolio.
Why It Matters
Widespread enterprise interest in DeepSeek models has surged due to cost efficiencies and competitive benchmark performance. NIST's risk findings establish an official baseline for enterprise risk managers, compliance teams, and security officers assessing foreign-developed open and proprietary models. Organizations deploying these models must review their risk posture against CAISI findings to ensure adequate safeguards and compliance with emerging standards.
Part of an ongoing development
SourceCAISI releases evaluation of DeepSeek AI models
National Institute of Standards and Technology's Center for AI Standards and Innovation (CAISI) has released an official evaluation of multiple artificial intelligence models from China-based DeepSeek. This review provides an authoritative U.S. government technical analysis regarding the operational reliability and risk profiles of DeepSeek's widely used model portfolio. Claims are as reported; this summary makes no determination about accuracy or significance.
- Confidence
- Low confidence
- Corroboration
- Unconfirmed
What we know
- Product:DeepSeek models
- Availability:General availability
- Available in:United States
- Organization:Center for AI Standards and Innovation (CAISI)
Organizations & Entities
Related Intelligence
- DevelopmentDevelopingAlso involving China
Mistral AI raises €3B in Series D funding
Mistral AI announced that it has secured €3 billion in a Series D funding round at a post-money valuation exceeding €21 billion. Claims are as reported; this summary makes no determination about accuracy or significance.
2 independent sources - DevelopmentNewAlso involving China
DeepSeek plans Huawei chip data center in Inner Mongolia
The Decoder reports that DeepSeek plans to establish a data center in Inner Mongolia equipped with 160,000 Huawei Ascend-950DT chips dedicated strictly to inference rather than training. If completed, the facility would represent the largest known Huawei AI processor cluster, though reported production bottlenecks mean Huawei may take over a year to deliver the hardware. Claims are as reported; this summary makes no determination about accuracy or significance.
1 reporting source - DevelopmentNewAlso involving China
Nvidia optimizes hardware support for Chinese open-weights AI models
Nvidia is actively optimizing its hardware architecture to support open-weights AI models developed in China, specifically DeepSeek and Alibaba's Qwen family. Concurrently, the chipmaker has issued warnings that prospective United States regulatory actions aimed at curbing the adoption or distribution of Chinese-origin AI models could negatively impact its business. Claims are as reported; this summary makes no determination about accuracy or significance.
1 reporting source - DevelopmentNewAlso involving China
Foreign adversaries illicitly access Anthropic AI systems for model distillation
CNBC reports that foreign adversaries are illicitly accessing American artificial intelligence systems to perform model distillation, creating cheaper copycat models and shifting unauthorized activity to the dark web amid heightened concerns involving Anthropic and China. Claims are as reported; this summary makes no determination about accuracy or significance.
1 reporting source