Move Over JARVIS, Meet OSCAR

Open-Source, Cloud-Based, AI-Enabled Reporting for the Intelligence Community

Vast new sources of publicly available information combined with Artificial Intelligence/Machine Learning (AI/ML) capabilities have revolutionized the way we think about data and insight. With the help of unclassified cloud capability, the intelligence community (IC) can reap the benefits of this revolution. Treating these new forms of open-source intelligence as an afterthought, however, will blind the IC to critical information—information the United States’ adversaries have already embraced and employed.

This revolution for the U.S. Intelligence Community is within reach. In the near term, a basic form of Tony Stark’s JARVIS could assist analysts and operators in a wide range of duties—from machine vision processing of millions of images to natural language processing of millions of pages of documents. This paper hypothesizes an open-source, cloud-based, AI-enabled reporting capability for the intelligence community named “OSCAR.” A capability like OSCAR could save an IC analyst as much as 364 hours, or more than 45 working days, a year, by some estimates.

So why has the IC been slow to embrace these advantages? This report breaks down the elements of ingrained culture, security practices, and policy decisions that hold back adoption of a system like OSCAR. In the “Possibilities Ahead” section, it looks to near-, mid-, and long-term opportunities for the IC to adopt OSCAR and accelerate intelligence work. Finally, it puts forth a slate of actionable recommendations that will break the logjam and allow the IC to obtain and deliver intelligence from anywhere.

This brief is made possible by support from Amazon Web Services.

Emily Harding
Director, Intelligence, National Security, and Technology Program and Deputy Director, International Security Program