Seven Republican senators have urged Commerce Secretary Howard Lutnick to identify and evaluate potential security vulnerabilities posed by Chinese open-source AI models like DeepSeek.Citing news articles on DeepSeek Artificial Intelligence (AI), the lawmakers pointed to concerns around DeepSeek having “willingly provided and will likely continue to provide support to China’s military and intelligence operations.”
These allegations are deeply troubling. DeepSeek’s R1’s model release in late January demonstrated the aptitude of People’s Republic of China (PRC) national AI talent and the progress their home-grown models have made relative to leading US products, the letter said.
R1 offers open-source access to its model weights, meaning that software engineers are able to modify portions of its code in order to further refine outputs to meet specific purposes or complete specialised tasks.Shortly after R1’s release, Wiz Research found a publicly accessible database belonging to DeepSeek, which allowed full control over database operations, including the ability to access internal data. Furthermore, R1 has been found to produce potentially harmful content at higher rates than peer American models, the senators said.
These allegations are deeply troubling. DeepSeek’s R1’s model release in late January demonstrated the aptitude of People’s Republic of China (PRC) national AI talent and the progress their home-grown models have made relative to leading US products, the letter said.
R1 offers open-source access to its model weights, meaning that software engineers are able to modify portions of its code in order to further refine outputs to meet specific purposes or complete specialised tasks.Shortly after R1’s release, Wiz Research found a publicly accessible database belonging to DeepSeek, which allowed full control over database operations, including the ability to access internal data. Furthermore, R1 has been found to produce potentially harmful content at higher rates than peer American models, the senators said.
The lawmakers urged the department to identify and act against four areas of concern.
- Any threats that may arise from data collected by such applications being fed back to servers located in China.
- Any information regarding verifiable or likely cases of DeepSeek and other Chinese open-source models feeding American personal or enterprise data to the Chinese People’s Liberation Army (PLA) or to companies known to be part of China’s military-industrial complex or surveillance apparatus.
- Any findings related to how Chinese open-source models may have improperly accessed export-controlled semiconductors or violated the use terms of US models to advance their capabilities.
- Work with the Center for AI Standards and Innovation (CAISI), and other relevant agencies with expertise in cybersecurity, data protection, and national security to investigate how US businesses and persons could be harmed by AI applications built atop R1 and other Chinese open-source models.
The senators also referred to previous similar action by actions against the widespread adoption of technologies from China, namely the Congress funded a multi-billion-dollar program to remove Huawei telecommunications hardware from American networks.
Finally, the letter urged the commerce department to provide a briefing to members of Congress and relevant Congressional Committees on findings as soon as practical.
