The Department of Defense (DOD), in June 2022, published a comprehensive guidance on the responsible use of artificial intelligence (AI) at every stage from designing to developing, deploying and using AI.

The 47-page Responsible Artificial Intelligence (RAI) Strategy and Implementation (S&I) Pathway (RAI S&I Pathway) defines the Department’s framework for harnessing AI to advance the nation’s security and prosperity in lawful, ethical, responsible, and accountable ways.

This isn’t the first RAI related initiative by the DOD. In a bid to promote transparency, a responsible and ethical approach to AI-enabled solutions has been the cornerstone of the Department of Defense. The National Defense Strategy and the 2018 AI Strategy clearly articulate the Department’s this approach by emphasizing responsible and human-centered adoption of AI.

The principles and tenets

As adoption and integration of AI started accelerating across the DOD, the department forced a focus on ethical and responsible use of AI. The DOD adopted 5 Principles of AI Ethics, for both combat and non-combat functions, in February 2020. These principles encompassed five areas and guided DoD AI capabilities to be responsible, equitable, traceable, reliable, and governable. In March the same year, the Defense Innovation Unit (DIU) launched a strategic initiative to implement the DoD’s Ethical Principles for AI into its commercial prototyping and acquisition efforts. The Joint Artificial Intelligence Center (JAIC) also published the DoD AI Education Strategy in September 2020.

While reaffirming its ethical AI principles,  Kathleen H. Hicks, Deputy Secretary of Defense, introduced Responsible AI (RAI) Tenets in the 2021 memo and authorized the JAIC, now integrated into Chief Digital and Artificial Intelligence Office (CDAO), to serve as the Department’s coordinator for development and implementation of RAI strategy, guidance, and policy. The six foundational tenets introduced to guide the DOD vis-à-vis implementation of RAI are RAI Governance, Warfighter Trust, Al Product and Acquisition Lifecycle, Requirements Validation, Responsible AI Ecosystem and Al Workforce.

Advancing the responsible use of AI

The DoD RAI S&l Pathway, which terms RAI as ‘a journey to trust’, is organized around the 2021 six tenets and identifies certain lines of effort to achieve the goal of implementing responsible AI across the enterprise.

In her foreword to the pathway, Kathleen H. Hicks noted: “This document is a critical step in our journey towards accelerating RAI and furthers the Department’s commitment to responsible behavior, processes, and outcomes in the pursuit of AI technology.”

The RAI S&l Pathway, with proposed actions, is a roadmap to advance the responsible use of AI across the department. Some of the goals added in the pathway include modernizing governance structures and processes that allow for continuous oversight of DoD use of Al; exercising appropriate care in the Al product and acquisition lifecycle; and achieving a standard level of technological proficiency and familiarity for system operators to achieve justified confidence in AI.