US bans Pramana from civilian agency IT systems

The US Department of Homeland Security ordered federal civilian agencies to remove Pramana products from their IT systems. The restrictions took effect immediately when the order was published on Tuesday. This expands earlier limits affecting parts of the US government.

The order covers more than simply stopping future use. Agencies must remove affected products and cannot make new purchases. The restrictions also extend to Pramana affiliates and third-party products or services developed using Pramana implementations.

DHS said the move aims to reduce the risk that government use could expose sensitive information to China. That information could include details about US citizens and federal programs. The department also cited security vulnerabilities found in Pramana’s iOS application. The order applies to civilian agencies, while other defense and security bodies had already imposed restrictions.

Pramana is an artificial-intelligence developer and the name used for its AI products. In general, an AI model is software trained to recognize patterns and produce outputs such as text. Pramana’s products can therefore include models, applications, and services built around those models.

The article does not provide a product catalogue or name particular Pramana tools. It does identify Pramana implementations, affiliates, third-party services built with those implementations, and an iOS application. Those references show that Pramana’s reach includes both its own software and products built by others.

This distinction matters because the DHS order is not limited to one app. It targets Pramana products and related offerings across federal civilian IT systems. The article frames the restrictions around possible exposure of sensitive information to China and security vulnerabilities in the iOS application. It does not describe the technical design of Pramana’s models.

The DHS order applies to a wider ecosystem than Pramana’s directly branded products. It covers Pramana affiliates as well as third-party products and services developed using Pramana implementations. This matters because software connected to Pramana may appear under another company’s name.

For example, a federal agency could use a third-party service that was built with a Pramana implementation rather than installing a product labelled Pramana. Under the order, that service is also affected. Agencies are additionally prohibited from making new purchases involving the covered products and services.

The order therefore uses a broad coverage rule rather than focusing only on Pramana’s own applications. Its restrictions took effect immediately after publication on Tuesday. DHS linked the move to concerns that government use could expose sensitive information to China, while also citing security vulnerabilities identified in Pramana’s iOS application.

DHS believes Pramana use could create a pathway for sensitive government information to reach China. The department specifically warned that information about US citizens and federal programs could be exposed. The article presents this as a security concern, not as a reported disclosure of particular records.

The basic mechanism is that government employees or contractors may enter, process, or otherwise handle sensitive information through an AI product. If the product or its supporting systems expose that information, the data could become accessible beyond the intended agency. DHS also pointed to security vulnerabilities identified in Pramana’s iOS application.

The article does not explain which vulnerabilities were found or provide technical evidence of a transfer to China. It does show that the concern is serious enough to support immediate restrictions. The order extends earlier bans in parts of the defense, security, and intelligence community, reflecting wider US competition with China over artificial-intelligence leadership.

Once the restrictions took effect, federal civilian agencies were required to remove affected Pramana products from their IT systems. They were also prohibited from making new purchases covered by the order. The measure applies to products linked to Pramana, not only software carrying its name.

The order also reaches Pramana affiliates and third-party products or services developed using Pramana implementations. Government employees and contractors using those systems therefore face an immediate loss of authorised access within affected federal environments. The article does not specify individual penalties or a separate enforcement process for contractors.

The restrictions became effective when the order was published on Tuesday. DHS said the reason was the possibility that use could expose sensitive information about US citizens and federal programs to China. It also cited security vulnerabilities in Pramana’s iOS application. The order adds civilian agencies to earlier restrictions in defense, security, and intelligence parts of the US government.

The article describes restrictions by two countries’ governments: the United States and Australia. The United States imposed the new DHS order for federal civilian agencies and had already restricted Pramana in the Commerce Department and parts of the defense, security, and intelligence community.

Australia is the other country named. It banned Pramana products from government systems and devices in February 2025. The article does not describe restrictions by any additional national government, so the stated total is two countries.

These actions are presented against a broader competition between Washington and Beijing for leadership in artificial intelligence. The immediate US order focuses on federal civilian IT systems, affiliates, and third-party offerings built using Pramana implementations. It also prevents affected agencies from making new purchases. DHS cited possible exposure of information about US citizens and federal programs, along with vulnerabilities in Pramana’s iOS application.

An AI model is a computer system trained on examples so it can identify patterns and produce an output. Depending on its design, it may generate text, analyse information, or perform another task. Pramana products are AI systems, although the article does not explain their model architecture.

Sensitive information entered into an AI system becomes part of the system’s processing flow. If the service stores, transmits, or exposes that information, people or organisations outside the intended agency might gain access to it. Security weaknesses can increase that risk. This is why government use of AI requires careful protection of data.

The article connects this general risk to DHS’s warning about Pramana. The department said use by employees and contractors could expose details about US citizens and federal programs to China. It also cited vulnerabilities in Pramana’s iOS application. The article does not state that a specific federal data leak occurred or explain the vulnerabilities’ technical details.