The Department of Defense’s race to operationalize artificial intelligence (AI) has entered a new phase. The fiscal 2026 National Defense Authorization Act (NDAA) allocated significant funding to accelerate AI adoption.
Building on this momentum, the DOD released its AI acceleration strategy, signaling a shift from experimentation to operational AI at scale. The guidance focuses on “establish[ing] a new AI execution standard for the entire department.”
The White House’s goal is clear. It is to equip the armed forces with next-generation AI capabilities as quickly as possible to establish “military AI dominance.” Failure to meet this goal results in battlefield vulnerability and a strategic disadvantage.
Achieving this ambitious goal requires tackling age-old problems, including data readiness, standardization and system security. If done right, the DOD will be well-poised to support American warfighting edge and future intelligence capabilities.
Legacy systems lagging
The DOD has struggled to scale AI efforts, largely remaining in isolated test phases. The AI acceleration strategy aims to change this by providing guidance on deployment and embedding AI into core mission workflows and planning.
However, several longstanding challenges stand in the way:
- Critical data within the DOD remains siloed, inconsistently labeled and inaccessible in real time. Many legacy systems were not built to support AI integration, which complicates ingestion.
- Secure architectures are evolving. Deployment at scale requires real-time data sharing, yet barriers between classification levels, delays in edge environments and cybersecurity risks all complicate deployment.
- Governance frameworks have not caught up with the current AI acquisition and deployment landscape. Accountability for AI outcomes is unclear, ethical standards vary and approval processes for AI decisions are fragmented.
Addressing these challenges will require a department-wide focus on the gaps that matter most.
Breaking the bottlenecks
There are no quick fixes for these decades-old problems. The solutions must be rolled out in phases, with the first phase focusing on architecture.
The DOD must establish an enterprise-like architecture built on common standards. Without this baseline, breaking down silos will be impossible, and AI tools will remain in the isolated test phase. AI needs secure, real-time data across domains and classification levels to operate effectively. Without this infrastructure, nothing can move forward.
In conjunction, the DOD must invest in centralized AI governance boards with operational authority. These can help prevent duplicative efforts, determine and enforce AI standards and ensure any AI investments align with the department’s strategy and vision.
The Pentagon Chief Digital and Artificial Intelligence Office’s Open Data and Applications Government-owned Interoperable Repositories (DAGIR) program is a powerful example of this effort in motion. With the goal of building a centralized platform for analytics and AI, this program breaks down silos and ensures AI tools are deployed consistently and securely.
Once these baselines have been established, another vital piece is ensuring operators, on-the-ground warfighters and leaders can leverage AI as a decision advantage.
Closing the human gap
While investments in technology and guidance are necessary, investments in the people who use them are equally important. Not every employee needs to be an AI expert, but they must understand what AI does and how to assess risk. Without baseline literacy, AI adoption will either stall or accelerate without proper oversight.
The DOD should invest in mandatory AI literacy training at all leadership levels. As AI becomes more integrated into processes, inter-departmental teams, including operators and acquisition professionals, should be formed in tandem. This will reduce the gap between the test phase and full deployment.
DOD can also retain AI talent through career pathways and incentives that reward AI fluency. Finally, leadership must acknowledge that in time-sensitive scenarios (e.g., missile defense), humans cannot evaluate every micro-decision. Rather than requiring human approval before any action, the better approach is to allow systems to act autonomously within defined parameters, with humans supervising only critical decisions.
Ultimately, successful alignment with the DOD’s AI Acceleration Strategy will see AI embedded in operational planning cycles, autonomous coordination and hybrid command structures. By addressing these challenges, the DOD can create a strategic “AI-first warfighting force” well-equipped to defend national interests in the future.
