news | 11 October 2022

Discovering Networks of Adversarial Influence Using Publicly and Commercial Available Information To Reveal Investment Trends and Strategies by Adversaries

DIU successfully transitions prototypes to counter foreign malign influence

MOUNTAIN VIEW, CA (October 11, 2022)—The United States faces a growing threat of state-based adversaries and opponents who leverage diplomatic, economic, military, and political activities to expand their global influence. Authoritarian governments and rogue states alike enlist elite networks of government officials, commercial entrepreneurs, and business executives to broker relationships and business opportunities that expand their foreign policy objectives and priorities abroad. Much of this activity occurs in the open and is discoverable through publicly- or commercially-available information (PAI/CAI).

Traditional analytic tools and models are manually intensive and oftentimes insufficient in terms of identifying the degree to which adversary activities are discoverable in open data.  In March 2020 and  September 2020, the Defense Innovation Unit (DIU) launched prototype projects with Accrete AI Government, LLC and Quantifind, Inc. respectively, to address the need for an innovative technological solution capable of discerning connections between the overseas activities of state-based adversaries, their internal and external investments, as well as their foreign policy objectives and strategies.

quantifind square image for website

Quantifind Knowledge Graph

These data sources reveal infrastructure investments by adversarial governments across multiple developing countries, the expansion of their banking and financial relationships in fragile capital markets, and their efforts to obscure these activities in seemingly mundane and routine ventures.

Each solution compiles news sources, blogs, social media, patent databases, academia, and other large volumes of open-source web-based content into an intelligent search platform where analysts and investigators can surface connections between individuals, organizations, and networks of interest.

The two solutions have since enabled the Department of Defense (DoD)  to rapidly utilize PAI/CAI data to identify and analyze patterns and relationships from opponent state activities. These solutions are presenting PAI/CAI data on adversarial state activities in seamless and visually compelling ways. The AI methodology delivered by Quantifind and Accrete provides tools to “extract,  structure, and visualize relevant data points, resolve what is true, and isolate the “so what” in order to bring timely recommendations to decision makers,” said Jaime Fitzgibbon, DIU AI program manager. The AI models developed through both solutions are continuously informed and validated by DoD analysts, to ensure that they reveal and represent the full range of risks associated with malign adversary networks.

In September 2022, both Accrete and Quantifind were awarded production contracts by the DoD that expand access to both solutions across the Department. Together, these solutions allow the (DoD) to both understand and respond to adversarial risks and attacks on US security interests with both the speed and urgency required to protect U.S. national security.

Quantifind and Accrete also applied DIU’s Responsible AI (RAI) Guidelines during the planning, development, and deployment phases of their solutions. These RAI guidelines ensure that the commercial technology delivered to the Department of Defense, through DIU, is effectively examined, tested, and validated to ensure compliance with the Department’s own Ethical Principles for AI.