The authors review current U.S. Department of Defense generative artificial intelligence acquisition efforts (focusing on influence activities) and provide recommendations for cost-effective acquisition and development.

Acquiring Generative Artificial Intelligence to Improve U.S. Department of Defense Influence Activities

William Marcellino, Jonathan Welch, Brittany Clayton, Stephen Webber, Thomas Goode

ResearchPublished Jul 22, 2025

Generative artificial intelligence (AI) presents opportunities for scaling and automation of tasks and activities related to influence activities conducted by the U.S. Department of Defense (DoD). DoD needs to rapidly acquire and employ generative AI capabilities to stay ahead of adversaries; however, ad hoc, bottom-up efforts to operationalize this technology create inefficiencies in acquisition and development related to common services and platforms, human capital and cross-functional teams, and contracting.

The authors of this report conducted a review of current DoD generative AI acquisition efforts (focusing on influence activities). They interviewed 18 subject-matter experts from DoD, the private sector, and government research organizations to identify force requirements and commercially available AI capabilities. They also held an expert workshop with 24 participants from various DoD influence organizations to elicit their generative AI–relevant operational and tactical needs. Drawing on this analysis, the authors provide recommendations for cost-effective acquisition and development to take advantage of current and emerging capabilities.

Generative AI acquisition has meaningful differences from traditional hardware and software acquisition, so the services should identify appropriate organizations to manage AI acquisition. The Principal Information Operations Advisor (PIOA) should direct the Office of Information Operations Policy (OIOP) to coordinate with influence-tasked units; U.S. Special Operations Command (USSOCOM) and U.S. Cyber Command (USCYBERCOM); service influence, information, and information operations organizations; and operational units with influence responsibilities to foster collaboration.

Key Findings

  • To effectively compete and counter adversaries, DoD has a clear need to enable the influence community with generative AI, but there is a substantial lack of investment and unity of effort at present.
  • Generative AI can improve analysis, operational planning, and assessment of influence activities. However, generative AI technology is a tool, not the answer, for addressing these rapidly evolving challenges.
  • Effective generative AI acquisition will need a strategic, flexible approach and a sustainment process that covers the spectrum of enterprise to bespoke capabilities.
  • No enterprise-wide plan or strategy currently addresses generative AI implications or opportunities as they relate to influence activities or operations in the information environment.

Recommendations

  • OIOP should encourage the military services and USSOCOM and USCYBERCOM to take the following actions: Define the formal requirements for influence activities, encourage investment in generative AI capabilities for influence, foster collaboration among influence stakeholders, and coordinate with DoD enterprise AI agencies (e.g., the Chief Digital and Artificial Intelligence Office) to leverage common infrastructure for influence.
  • The services should identify appropriate organizations to implement the following actions: Leverage the suite of available acquisition strategies to enable flexibility across generative AI capabilities that range from broad, DoD-wide capabilities to bespoke, in-house–developed technologies; implement a formal process to define generative AI capability requirements (this would not fall under the purview of the acquisition community but should be coordinated at some level across organizations dedicated to influence so as to synchronize like requirements); increase the tempo for capability purchases or reassessment; and develop and coordinate sustainment strategies for generative AI capabilities across influence stakeholders.
  • PIOA and OIOP should develop guidance to enable effective and efficient adoption across the influence community, identify and invest in AI training and education opportunities, and develop guidelines (guardrails) to govern the use of AI-generated output in influence activities.

Topics

Document Details

Citation

Chicago Manual of Style

Marcellino, William, Jonathan Welch, Brittany Clayton, Stephen Webber, and Thomas Goode, Acquiring Generative Artificial Intelligence to Improve U.S. Department of Defense Influence Activities. Santa Monica, CA: RAND Corporation, 2025. https://www.rand.org/pubs/research_reports/RRA3157-1.html.
BibTeX RIS

Research conducted by

This research was sponsored by the Influence and Information Capabilities Subgroup of the Irregular Warfare Technical Support Directorate (IWTSD) in the Office of the Under Secretary of Defense for Policy and conducted within the Acquisition and Technology Policy Program of the RAND National Security Research Division (NSRD).

This publication is part of the RAND research report series. Research reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.