The U.S. Department of Defense uses artificial intelligence as part of a wider data-and-analytics program to improve battlespace awareness, planning, decision support, logistics, cyber operations and business processes. Public documents describe a mixture of existing capabilities, announced pilots and research projects—not one universal AI system and not proof that every listed use is deployed across the force.
What “using AI” means inside the Pentagon
DoD uses “AI” broadly. The term can include machine-learning models that classify or forecast, analytics that help people interpret large data sets, generative systems that produce or summarize content, and autonomy software that helps an uncrewed system perceive and act. The department’s public strategy treats all of these as parts of a larger data ecosystem rather than as a single product.
That distinction matters because official announcements describe different levels of maturity. A strategic goal is an intended outcome; a pilot is a bounded test; a competition can demonstrate a technique without putting it into a military network; and a fielded capability is one that has passed the department’s own acquisition, security and operational processes.
| Evidence level | What it tells you | What it does not prove |
|---|---|---|
| Strategic outcome | What DoD wants data and AI to improve. | That a particular model or system is already deployed. |
| Announced pilot | Where the department is directing an initial experiment or capability effort. | Department-wide availability, operational effectiveness or final spending. |
| Research or competition | That a technique is being explored under controlled conditions. | That the technique is used autonomously in live operations. |
| Fielded system | A capability has entered operational use, if the announcement explicitly says so. | That every service, command or mission uses it in the same way. |
The most reliable way to read a claim is therefore to ask which level it describes, who made the statement, when it was made and whether the source names a deployed system or only an objective.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →#1 Best Overall
DoD’s five strategic outcomes for AI and analytics
The November 2023 release of the Data, Analytics, and Artificial Intelligence Adoption Strategy and the strategy document organize adoption around five outcomes:
| Outcome | Practical meaning |
|---|---|
| Superior battlespace awareness and understanding | Combine and interpret information so commanders and staff can form a clearer, faster picture of changing conditions. |
| Adaptive force planning and application | Use data and models to evaluate options, adjust plans and apply forces as circumstances change. |
| Fast, precise, resilient kill chains | Improve the connected process that moves from sensing and understanding a situation to deciding and acting, while making that process faster and more robust. |
| Resilient sustainment support | Help maintain people, equipment and supply networks by improving forecasting, prioritization and response to disruption. |
| Efficient enterprise business operations | Apply analytics and automation to administrative, financial, human-resources and other support work, not only to combat missions. |
The strategy presents these outcomes as organizational goals. It does not provide a public list showing that every outcome has been achieved throughout DoD. Its approach is to build the conditions in which military and civilian personnel can use data and AI repeatedly and safely.
Where DoD announced generative-AI pilots
On December 11, 2024, the Chief Digital and Artificial Intelligence Office and the Defense Innovation Unit announced an initial set of generative-AI pilot focus areas through the AI Rapid Capabilities Cell. The announcement said the cell would initially receive approximately $100 million across fiscal years 2024 and 2025 for pilot efforts, foundational infrastructure and tools. That was planned resourcing announced at that date, not a final expenditure total.
| Pilot area | What the area covers |
|---|---|
| Command and control and decision support | Tools that help leaders and staffs organize information, understand options and make decisions. |
| Operational planning | Applications that assist with developing, comparing or updating plans. |
| Logistics | Use of generative systems in supply, movement, maintenance and other sustainment workflows. |
| Weapons development and testing | Support for engineering, evaluation and test activities around weapon systems. |
| Uncrewed and autonomous systems | AI-related capabilities for systems that operate with reduced or remote human control. |
| Intelligence activities | Processing and analysis that can help intelligence professionals handle large or complex information sets. |
| Information operations | Support for activities involving information, audiences and the information environment. |
| Cyber operations | Use of generative tools in defensive and offensive cyber mission processes, subject to applicable controls. |
These are announced pilot categories, not a claim that a generative model is operating across every listed mission. A pilot can be limited by classification, network access, data quality, security testing, model performance or the availability of trained personnel.
Free tools Windows power users keep installed
One-click scans. No signup required.
How AI fits specific defense functions
Battlespace awareness and intelligence
AI can help personnel find patterns, correlate information from different sources and flag items for human review. The strategic objective is better understanding of the battlespace, while the public documents do not identify one universal model or provide a department-wide accuracy figure. Intelligence use remains dependent on the quality, timeliness and classification of the underlying data.
Rank #2
Command, control and decision support
Decision-support systems can present relevant information, generate summaries, compare courses of action or identify uncertainty. The stated purpose is to improve the speed and quality of human decisions. DoD’s strategy frames AI as a tool for people rather than a substitute for command authority.
Planning and kill chains
“Kill chain” in the strategy refers to the connected process by which forces sense, understand, decide and act. AI and analytics may support parts of that process, but the strategy’s phrase does not by itself establish that an algorithm independently chooses targets or authorizes the use of force. Rules of engagement, legal review, command responsibility and technical assurance remain separate requirements.
Logistics and sustainment
Supply and maintenance networks produce large volumes of records. Analytics can help forecast demand, identify bottlenecks, prioritize scarce parts or anticipate maintenance needs. The strategic outcome is resilient sustainment support; public strategy material does not give a single accuracy rate, platform list or universal deployment date for these functions.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchUncrewed and autonomous systems
Autonomy can cover perception, navigation, route planning or other bounded tasks in an uncrewed system. The 2024 pilot announcement places uncrewed and autonomous systems among the areas to explore with generative AI, but a pilot label should not be read as evidence that weapons operate without human direction in ordinary service.
Cyber and information operations
AI can assist analysts and operators with finding anomalies, triaging alerts, drafting material or examining software. The announced pilot list identifies cyber operations and information operations as focus areas; it does not publish a blanket authorization for autonomous cyber action or automated influence campaigns.
Enterprise business operations
DoD is also a very large employer and administrator. Its strategy explicitly includes enterprise business operations, so AI adoption can involve noncombat work such as records processing, workflow support and other administrative tasks. This is an important part of the department’s stated plan, even though it receives less public attention than weapons or intelligence.
A concrete program example: DARPA’s AI Cyber Challenge
The Defense Advanced Research Projects Agency’s AI Cyber Challenge is a specific, publicly described demonstration of how DoD-backed work applies generative AI. According to the March 27, 2024 DoD account, the challenge uses generative-AI technologies to automatically find and fix vulnerabilities in open-source software, including software that supports critical infrastructure.
The challenge is a DARPA competition and research activity. It shows that automated vulnerability discovery and repair are being explored; it does not establish that deployed DoD networks autonomously patch all such software. Real-world use would still require validation, secure integration, authorization and monitoring.
The foundations DoD says AI needs
The 2023 strategy places data, infrastructure, people and governance ahead of ambitious applications. Its “hierarchy of needs” identifies the following foundations:
- Interoperable infrastructure: systems must exchange data and services rather than trapping information in isolated environments.
- Data quality and foundational data management: models are only as useful as the data’s accuracy, completeness, provenance, timeliness and accessibility.
- Digital talent: military and civilian personnel need skills to build, operate, evaluate and challenge AI systems.
- Analytics and measurable impact: organizations need metrics that show whether a capability improves a mission instead of merely producing an impressive demonstration.
- Assurance: testing and monitoring must address reliability, security, robustness and performance under changing conditions.
- Governance: authorities, accountability and decision rights have to be clear across services, commands and civilian organizations.
- Responsible use: AI must be developed and employed within applicable law, policy and ethical controls.
This foundation-first model explains why an AI announcement may discuss data catalogs, cloud or network architecture, workforce development and governance in the same breath as a model. The department is trying to make capabilities repeatable and supportable, not simply to purchase isolated demonstrations.
Responsible AI and the human role
Deputy Secretary of Defense Kathleen Hicks explained the motivation in the 2023 strategy release: “As we focused on integrating AI into our operations responsibly and at speed, our main reason for doing so has been straight forward: because it improves our decision advantage.” The emphasis is on improving people’s decisions, with responsibility retained by authorized personnel.
Craig Martell, then DoD chief digital and artificial intelligence officer, similarly said the strategy was not a list of a few supposedly decisive systems. Instead, it described an environment in which people could continuously deploy data, analytics and AI capabilities for enduring decision advantage. That framing places oversight, evaluation and organizational learning at the center of adoption.
DoD also lists a June 2024 Responsible Artificial Intelligence Strategy and Implementation Pathway. The available official description presents it as AI-specific guidance that complements the department’s existing framework. Claims about individual principles or implementation steps should be tied to the full document rather than inferred from its title.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How long has DoD been using AI?
The foreword to the 2023 adoption strategy states that DoD had been investing in AI and responsibly fielding data- and AI-enabled systems for more than 60 years. In a March 5, 2024 speech, Hicks referred to earlier examples ranging “from air defense networks to logistics planning tools.” These statements provide historical context, but they do not constitute a complete system-by-system chronology.
A separate statistic needs careful attribution. In a March 2024 DoD account, Matt Turek, deputy director of DARPA’s Information Innovation Office, said AI, machine learning and autonomy were used by “about 70% of DARPA’s programs in some form.” That is an attributed estimate about DARPA programs, not a DoD-wide percentage and not a measurement of operational deployment.
Best Value
How industry and government systems are connected
DoD’s May 30, 2024 Open DAGIR announcement describes a multi-vendor approach for integrating data platforms, development tools, services and applications. The department says the model is intended to preserve government ownership of government data while protecting companies’ intellectual property and allowing vendors to build applications that use approved government data.
That architecture addresses a practical problem: DoD needs to adopt useful capabilities without locking every mission into one proprietary stack. It also means that an announcement about an ecosystem or contract vehicle should not be mistaken for proof that a particular application is deployed in every command.
How to judge a claim that DoD “uses AI”
When a news release or documentary describes a military AI capability, use this checklist:
- Identify the mission: Is it intelligence, command support, logistics, cyber, autonomy, weapons testing or business administration?
- Identify the maturity: Does the source say fielded, operational, pilot, prototype, challenge or strategy?
- Locate the human decision: What does the system recommend, generate or automate, and who approves the consequential action?
- Check the data and network conditions: Does the claim depend on classified data, a particular sensor, connectivity or a controlled test environment?
- Check the date and authority: A 2024 pilot announcement and a 2023 strategy are evidence of intent at those dates, not guarantees of later deployment.
- Separate a percentage from its scope: “About 70%” applies to DARPA programs in some form, not to the entire Department of Defense.
What the 2026 strategy listing does—and does not—establish
An official listing shows a document titled Artificial Intelligence Strategy for the Department of War, dated January 9, 2026: the listed PDF. The listing alone does not establish the document’s detailed use cases, implementation steps or relationship to the 2023 DoD adoption strategy. Until its full text and institutional context are available for review, those details should not be presented as confirmed facts.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Bottom line
DoD uses AI to strengthen how people collect and understand information, plan and apply forces, connect sensing to action, sustain equipment and run a huge administrative enterprise. Its public record also includes generative-AI pilots and DARPA demonstrations aimed at cyber defense and autonomous capabilities. The department’s own strategy makes clear that reliable data, interoperable infrastructure, skilled personnel, governance, assurance and responsible human oversight are prerequisites. The evidence supports a department-wide push with a mix of fielded work, pilots and research—not a claim that one autonomous AI system runs U.S. defense operations.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
