Skip to main content

Artificial Intelligence: DOD Needs Department-Wide Guidance to Inform Acquisitions

GAO-23-105850 Published: Jun 29, 2023. Publicly Released: Jun 29, 2023.
Jump To:

Fast Facts

The Department of Defense is developing artificial intelligence capabilities—computer systems that can do tasks that normally require human intellect.

The private sector has been acquiring AI for years. Thirteen private companies told us about their AI acquisition practices. For example, some companies mentioned the importance of considering intellectual property and data rights when negotiating contracts for AI projects.

Although parts of DOD are already using AI, DOD hasn't issued department-wide AI acquisitions guidance needed to ensure consistency. We recommended it develop such guidance—considering private company practices as appropriate.

Illustration showing the Pentagon on the left overlaid with a network of dots and an outline of a woman's profile with the letters "AI" emblazoned on her.

Skip to Highlights

Highlights

What GAO Found

The Department of Defense (DOD) designated artificial intelligence (AI) a top modernization area and is allocating considerable spending to develop AI tools and capabilities. AI refers to computer systems designed to replicate a range of human functions and continually get better at their assigned tasks. DOD AI capabilities could be used in various ways, for example in identifying potential threats or targets on the battlefield.

GAO obtained information from 13 private sector companies about how they successfully acquire AI capabilities. Elements of these categories, shown below, are also reflected in GAO's June 2021 AI Accountability Framework report (GAO-21-519SP).

Categories of Factors Selected Companies Reported Considering When Acquiring Artificial Intelligence Capabilities

Categories of Factors Selected Companies Reported Considering When Acquiring Artificial Intelligence Capabilities

Although numerous entities across DOD are acquiring, developing, or already using AI, DOD has not issued department-wide guidance for how its components should approach acquiring AI. DOD is in the process of planning to develop such guidance, but it has not defined concrete plans and has no timeline to do so. The military services also lack AI acquisition-specific guidance, though military officials noted that such guidance would be helpful to navigate the AI acquisition process. Without department-wide and tailored service-level guidance, DOD is missing an opportunity to ensure that it is consistently acquiring AI capabilities in a manner that accounts for the unique challenges associated with AI.

Various DOD components and military services have individually developed or plan to develop their own informal AI acquisition resources. Some of these resources reflect key factors identified by private companies for AI acquisition. For example, DOD's Chief Digital and AI Officer oversees an AI marketplace known as Tradewind, which is designed to expedite the procurement of AI capabilities. Several Tradewind resources emphasize the need to consider intellectual property and data rights concerns when negotiating contracts for AI capabilities, a key factor identified by the companies GAO interviewed.

Why GAO Did This Study

DOD has begun to pursue increasingly advanced AI capabilities. DOD has historically struggled to acquire weapon systems software, and AI acquisitions pose additional challenges. In February 2022, GAO described the status of DOD's efforts to develop and acquire AI for weapon systems.

Senate Report 116-236 accompanying the National Defense Authorization Act for Fiscal Year 2021 includes a provision for GAO to review DOD's AI acquisition efforts. This is the second report in response to that provision. This report examines (1) key factors that selected private companies reported considering when acquiring AI capabilities, and (2) the extent to which DOD has department-wide AI acquisition guidance and how, if at all, this guidance reflects key factors identified by private sector companies.

GAO analyzed information provided by 13 private companies with expertise in designing, developing, and deploying AI systems in various sectors to determine the key factors. GAO also analyzed DOD documentation and compared it with the key factors, and interviewed DOD officials.

Recommendations

GAO is making four recommendations for DOD and the three military departments to develop guidance on acquiring AI capabilities, leveraging private company factors as appropriate. DOD concurred with the recommendations.

Recommendations for Executive Action

Agency Affected Recommendation Status
Department of Defense The Secretary of Defense should ensure that the Chief Digital and AI Officer, in conjunction with other DOD acquisition policy offices as appropriate, prioritize establishing department-wide AI acquisition guidance, including leveraging key private company factors, as appropriate. (Recommendation 1)
Open – Partially Addressed
DOD agreed with this recommendation. To address it, the Chief Digital and AI Officer (CDAO) developed the DOD Data, Analytics, and AI Adoption Strategy, released in November 2023, which includes guidance for DOD components on adopting and scaling AI capabilities, with a forthcoming decision aid framework appendix. DOD is also planning to pursue an iterative approach to identifying best practices for acquisition of data, analytics, and AI, including forthcoming information on DOD access to AI training data, guidelines for generative AI, a steering group on acquisition and contracting practices for business models, and establishing a contracting vehicle for acquiring enterprise solutions.
Department of the Army After DOD issues department-wide AI acquisition guidance, the Secretary of the Army should establish service-specific AI acquisition guidance that includes oversight processes and clear goals for these acquisitions, and leverages key private company factors, as appropriate. (Recommendation 2)
Open
The Army agreed with this recommendation. As of Fall 2023, it noted that it expects to update the AI and Autonomy Roadmap after CDAO publishes its AI strategy. The Army is also delivering a Unified Data Reference Architecture and Project Linchpin, the Army's operations-enabling program for AI/machine learning capabilities. As of September 2024, the Army indicated it expects to complete these actions in FY2025.
Department of the Navy After DOD issues department-wide AI acquisition guidance, the Secretary of the Navy should establish service-specific AI acquisition guidance that includes oversight processes and clear goals for these acquisitions, and leverages key private company factors, as appropriate. (Recommendation 3)
Open
The Navy agreed with this recommendation. As of Fall 2023, it noted that it expects to update AI acquisition guidance issued by the Secretary of the Navy where applicable within 90 days after CDAO issues its AI strategy. As of September 2024, the Navy has not provided further information on its timeline for implementation of this recommendation.
Department of the Air Force After DOD issues department-wide AI acquisition guidance, the Secretary of the Air Force should establish service-specific AI acquisition guidance that includes oversight processes and clear goals for these acquisitions, and leverages key private company factors, as appropriate. (Recommendation 4)
Open
The Air Force agreed with this recommendation. As of Fall 2023, it noted that it expects to address this recommendation by January 2026 by establishing Air Force-specific AI acquisition guidance after CDAO issues its AI strategy. We will continue to follow-up on the Air Force's plans to issue its AI acquisition guidance.

Full Report

GAO Contacts

Jon Ludwigson
Director
Contracting and National Security Acquisitions

Media Inquiries

Sarah Kaczmarek
Managing Director
Office of Public Affairs

Public Inquiries

Topics

Artificial intelligenceDefense acquisition programsIntellectual propertyPrivate sectorSoftwareSystems acquisitionWeapon systemsMilitary forcesCompliance oversightGovernment procurement