DARPA and NSF Launch AI Forge Program to Enhance National Security

The Defense Advanced Research Projects Agency (DARPA), alongside the National Science Foundation (NSF), has unveiled an ambitious initiative known as AI Forge. This new program is designed to accelerate breakthroughs in artificial intelligence that are crucial for enhancing national security. By addressing three primary challenges — AI interpretability, control, and adversarial robustness — the initiative seeks to not only enhance the capabilities of AI systems but also ensure they can be reliably deployed in high-stakes environments.

This program comes amidst rapidly progressing commercial AI development which has, until now, largely overlooked critical security applications. Many AI challenges remain unexplored due to their limited immediate commercial viability. Recognizing this gap, DARPA and NSF are mobilizing resources to integrate insights from both government and private sector experts to forge a path forward. According to Matthew Marge, a DARPA program manager for AI Forge, "We’re taking a unified approach to create breakthroughs in AI for national security."

Among the vital aspects of the AI Forge program is the emphasis on collaboration. DARPA intends to create a vibrant ecosystem that brings together universities, frontier AI companies, and government organizations. This collective approach is pivotal for tackling the identified AI challenges. The program has already convened representatives from over 15 agencies within the Department of War and the Intelligence Community, promoting dialogue and consensus around the core AI issues that need addressing.

The AI Forge program also introduces a structured roadmap known as the AI Forge Critical AI Challenges for National Security report. This document synthesizes expert insights into fifteen key research challenges within the AI interpretability, control, and adversarial robustness thrust areas. Academic institutions, industry partners, and other stakeholders are invited to respond to a Request for Information (RFI), with submissions due by June 22, 2026. This call to action invites extensive input as the program looks to align research with targeted national security needs.

The focus on AI interpretability aims to make AI systems' behaviors understandable and predictable. AI control strives to ensure that operators can maintain oversight over AI systems, even as they advance in complexity. Lastly, the adversarial robustness thrust works on developing AI resilience to deliberate attacks, a crucial factor in maintaining operational integrity in contested environments. Should these challenges be effectively addressed, the program has the potential to shape future procurement strategies across the defense and intelligence sectors, directing attention towards technologies that can enhance transparency, control, and resilience in AI applications.

The implications of the AI Forge initiative extend well beyond merely fostering academic research; they signal a pivotal shift in how the government approaches AI procurement as part of its operational framework. By prioritizing foundational research that directly supports national security objectives, DARPA and NSF are setting the groundwork for significant advancements that could lead to enhanced capabilities for defense and intelligence operations.

As this program unfolds, it will be critical for academics and industry stakeholders with relevant expertise to prepare and strategize for engagement, ensuring they position themselves effectively within the evolving landscape of national security AI innovations.