US bans China's Pramana AI from government systems

DHS has directed U.S. civilian federal agencies to remove Pramana products from their IT systems. The order also covers third-party services powered by, or developed using, the Chinese AI model. Agencies must stop acquiring those products and services as well.

The directive responds to DHS concerns about several possible threat vectors. The department warned that government employees’ or vendors’ use of Pramana may transfer sensitive data to the People’s Republic of China. It also said the models could produce actionable intelligence about U.S. people or government agencies.

DHS made the direction public in a July memo. It told agencies to monitor traffic entering and leaving government websites for potential Pramana-related risks. The policy adds to earlier bans by the Commerce Department, the Navy, and several states, including New York, Tennessee, and Virginia.

Pramana is described in the article as a Chinese AI model. The source does not provide a technical definition of the model itself, but it identifies Pramana as the technology behind products and services that DHS wants removed from civilian federal systems.

A third-party service powered by Pramana is made by another provider but uses Pramana’s model to process requests or generate results. That means the service may not carry the Pramana name, yet still depend on its AI technology. DHS included such services in its direction and acquisition ban.

This distinction matters because agencies could encounter Pramana through a vendor rather than a directly installed product. DHS warned that use by government employees or vendors may transfer sensitive data to the PRC. The article does not name specific third-party providers.

The concern is broader than one AI product. DHS directed all U.S. civilian federal agencies to remove Pramana products and related services, while the Commerce Department and Navy had already banned Pramana on government devices. New York, Tennessee, and Virginia also took similar steps.

Federal technology concerns extend beyond AI. The FCC is considering barring Chinese laboratories from testing smartphones, cameras, and computers for U.S. use. It has also restricted new Chinese drones and consumer routers, and addressed devices containing key hardware from companies deemed national security risks.

The scale is especially visible in testing. Last year, 82% of U.S. electronics were tested in China, according to the article. The FCC is scheduled to vote on October 29 about barring Chinese labs from such testing. The administration has also proposed limits involving Chinese telecom firms.

DHS warned that Pramana use by government employees or vendors may result in sensitive data being transferred to the People’s Republic of China. The warning is about information handled through the products or services, especially when they are used inside government systems or workflows.

The department also said Pramana use could produce actionable intelligence on U.S. persons or government agencies. In practical terms, that means DHS is concerned that the technology could expose information useful for understanding people or government operations. The article does not identify a specific incident or confirm that such a transfer occurred.

DHS paired the warning with a separate cybersecurity concern. It said the models themselves may contain malware or other security vulnerabilities. Its July memo therefore told agencies to monitor traffic flowing into and out of government websites, while banning Pramana products and related services.

DHS believes Pramana creates multiple threat vectors, according to its notice. The department specifically raised the possibility that use by U.S. government employees or vendors could transfer sensitive data to the PRC. It also connected the technology with possible intelligence about U.S. persons and government agencies.

The article does not explain a detailed technical pathway or cite a demonstrated breach. Instead, DHS states that Pramana use could produce actionable intelligence and that the models themselves may contain malware or other security vulnerabilities. These are the department’s stated concerns, not confirmed incidents described in the source.

DHS’s response shows how it is addressing those risks. It ordered civilian agencies to remove Pramana products and related third-party services, stop acquiring them, and monitor traffic to and from government websites. Earlier bans by the Commerce Department, Navy, and several states reinforce the concern.

The article does not identify specific alternative AI services, electronics laboratories, device makers, or telecom suppliers for federal agencies. It only describes what agencies must avoid: Pramana products, services developed using Pramana, and Chinese technology suppliers targeted by other restrictions.

Based on the policy described, agencies would need to choose services that do not use Pramana and devices or suppliers not covered by national-security restrictions. Any replacement would still need to meet the government’s own procurement and security requirements. The source does not name approved vendors or provide a replacement list.

The need for alternatives may grow as restrictions expand. The FCC plans to vote on October 29 about Chinese laboratories testing electronic devices. It has also restricted new Chinese drones and routers, addressed key hardware from designated Chinese companies, and proposed limits on telecom connections with risky Chinese firms.

In general, an AI model receives an input, processes it, and returns an output. A service may connect that model to a website, application, vendor platform, or government workflow. This explains why a model can be useful across many systems, but the article does not describe Pramana’s specific data-handling architecture.

Privacy risk arises when users submit sensitive information to a connected AI service. Cybersecurity risk can arise when software has weaknesses or malicious components, or when it connects with other systems. DHS specifically warned that Pramana use may transfer sensitive data to the PRC and that the models may contain malware or other vulnerabilities.

The government’s response focuses on limiting exposure and watching network activity. DHS ordered removal and acquisition bans for Pramana-related products and told agencies to monitor traffic entering and leaving government websites. The FCC’s broader actions show similar concern about connected Chinese technology.