Abstract

Assessing military AI capabilities presents a number of challenges. It is difficult to precisely gauge how effectively AI-based systems will perform in an operational environment, a task made more difficult by the natural secrecy that militaries employ with regard to the capabilities of their systems. However, AI itself might provide a way to potentially conduct such assessments. This webinar will explore the potential for open-source AI tools to be deployed to assess military AI capabilities, and the associated risks and benefits.

This event is part of a series organised by the Military Transformations Programme on emerging issues in military AI.

Speaker

Ritwik Gupta is a Ph.D. candidate at the University of California, Berkeley, where he is a part of the Berkeley AI Research (BAIR) lab. His research primarily focuses on computer vision, specializing in representation learning and developing architectures for processing large images. He is also focused on policy issues surrounding the dual-use nature of AI in civilian-military contexts. Ritwik’s technical research is directly motivated his expertise in the fields of humanitarian assistance and disaster response, with his work in vision finding direct impact in societal problems such as assessing building damage after the 2023 Turkey-Syria earthquake, or detecting and interdicting criminals engaged in illegal activities on the ocean.

Ritwik is a Graduate Fellow with the Berkeley Risk and Security Lab, a Research Fellow at Berkeley’s Human Rights Center, and an AI Policy Fellow at the Center for Security in Politics. Ritwik previously led a research lab focused on AI for humanitarian assistance and disaster response at Carnegie Mellon University and currently serves as the Technical Director for Autonomy at the Defense Innovation Unit and as an Advisor to the FBI on AI and AI Policy.