The Pentagon is ramping up efforts to fortify its artificial intelligence (AI) systems against potential vulnerabilities that could be exploited by adversaries employing visual tricks or manipulated signals. Under the Guaranteeing AI Robustness Against Deception (GARD) program initiated in 2022, researchers are delving into the realm of “adversarial attacks” to safeguard AI systems from fatal misidentifications.
Recent studies have demonstrated how seemingly innocuous visual patterns can deceive AI algorithms, leading to critical misinterpretations of objects. For instance, an AI could erroneously identify a bus filled with passengers as a tank if the right “visual noise” is applied. Such errors could have dire consequences, particularly on the battlefield.
Concerns surrounding the Pentagon’s development of autonomous weapons have spurred the Department of Defense to update its AI development protocols, placing greater emphasis on responsible behavior and requiring approval for all deployed systems.
Despite modest funding, the GARD program has made significant strides in developing defenses against adversarial attacks. Tools developed through this program have been provided to the newly formed Defence Department’s Chief Digital and AI Office (CDAO).
However, advocacy groups remain wary, expressing concerns that AI-powered weapons could misinterpret situations and initiate attacks without proper cause, potentially leading to unintended escalations, particularly in volatile regions.
To address these challenges, the Pentagon is actively modernizing its arsenal with autonomous weapons while acknowledging the urgency of fortifying AI systems against vulnerabilities. The Defense Advanced Research Projects Agency (DARPA) announced that researchers from Two Six Technologies, IBM, MITRE, University of Chicago, and Google Research have developed a range of resources, including virtual testbeds, toolboxes, benchmarking datasets, and training materials, to aid in the defense against adversarial attacks.
These efforts underscore the ongoing commitment to ensuring the responsible development and deployment of AI technology in defense applications, safeguarding against potential threats and vulnerabilities in an ever-evolving landscape of warfare.
Read Now:South Korean Scientists Achieve Breakthrough in Nuclear Fusion with KSTAR
