DARPA has transitioned defensive AI capabilities to the Chief Digital and Artificial Intelligence Office from its GARD program to protect AI systems from adversarial attacks and deception. The CDAO will distribute these techniques across the Department of Defense to enhance national security measures.
DARPA Transfers AI Defense Technology to Military’s AI Office
The Defense Advanced Research Projects Agency (DARPA) has successfully transitioned new defensive capabilities to the Chief Digital and Artificial Intelligence Office (CDAO), originating from its Guaranteeing AI Robustness Against Deception (GARD) program. This initiative aims to protect AI systems from adversarial attacks by developing techniques to counter deceptive practices, such as noise patterns in sensor data that can cause AI misclassification.
Matt Turek, Deputy Director of DARPA’s Information Innovation Office, highlighted these developments during a virtual event hosted by the Center for Strategic and International Studies. The GARD program collaborates with industry partners to create algorithms that prevent these types of attacks, ensuring AI systems can operate securely even in adversarial environments.
The CDAO, established in 2022, will facilitate the distribution of these capabilities across the Department of Defense (DoD). DARPA’s multi-pronged strategy also involves sharing new algorithms and tools with external researchers and organizations such as Google, enhancing overall defenses in AI and machine learning.
For fiscal year 2024, the Pentagon requested $10 million for the GARD program, which is scheduled to conclude thereafter. DARPA continues to innovate in the AI sector, dedicating $2 billion since 2018 through its AI Next campaign and proposing $310 million for the AI Forward initiative in the upcoming fiscal budget. These efforts aim to develop trustworthy AI technology that meets national security needs effectively and ethically.
