DARPA Selects Xint to Apply AI for Securing Military Messaging Applications

DARPA Partnership for Messaging Security

The Defense Advanced Research Projects Agency (DARPA) has chosen Xint to conduct research into autonomous AI application security. This research will focus on deep analyses of messaging applications, both internally and externally developed, that are utilized across the Department of War. The objective is to enhance the security of these critical communication tools.

Background on Xint's Selection

Xint, which originated from Theori, was selected by DARPA after being one of three winners in the Artificial Intelligence Cyber Challenge (AIxCC). This competition was a two-year, $29.5 million initiative. Xint's role now involves analyzing the source code of messaging applications, including those like Signal and open-source projects. It will also analyze compiled binaries, utilizing a new service set to launch in September 2026, which assesses software supply chain risk in compiled code.

Addressing Unique Messaging App Vulnerabilities

Andrew Wesie, CTO and co-founder at Xint, stated that messaging applications are unique because an attacker only needs read-only access to compromise their core function. He highlighted that third-party SDKs and libraries embedded in these apps can introduce hidden data risks, potentially exposing user location or personally identifiable information during sensitive communications without the knowledge of the user or developer.

DARPA's goal is to protect Department of War messaging from external, particularly foreign, eavesdropping. Xint's technology uses LLM models to examine all source code and reverse engineer binaries as needed. For example, in Android messaging, Xint would analyze the app, the Linux kernel, and other components between the app and the kernel.