The Department of Defense is making a bold investment in the future, dedicating $139.9 million to its Chief Digital and Artificial Intelligence Office for fiscal year 2025. This funding signals a strong push to accelerate AI and machine learning across military operations. Here’s a look at the key priorities and ambitious goals shaping the CDAO’s mission for the year ahead.
If you’re a member of government or industry who needs to know what the latest AI focus areas are in the government and military, look no further than the Potomac Officers Club’s 2025 AI Summit. On March 20, POC will gather the foremost experts in AI/ML for a day of informative and compelling keynotes and panel sessions, with many built-in networking sessions and Q&A opportunities. Tickets are going fast — register now!
Table of Contents
AI/ML Scaffolding Initiatives
The CDAO continues to expand its AI/ML scaffolding efforts to ensure robust integration of innovative technologies into complex DOD systems. Key initiatives include:
AI Testing and Evaluation and Assurance Case Best Practices
- Focus: Development of guidance for T&E of high-consequence systems and integration of AI into operator workflows.
- Outputs:
- AI Operational Testing and System Integration Testing Frameworks.
- DOD large language model benchmarks and usage guides.
- Comprehensive assurance frameworks for novel AI technologies, emphasizing their safe usage and understanding of associated risks.
DREAM Program
- Goals:
- Develop AI/ML reference architecture models tailored to DOD mission needs.
- Facilitate collaboration with industry partners and CTO peers to exchange strategic insights.
- Advance vendor engagement to align technological trends with department requirements.
- Conduct experiments and surveys to refine data management policies and frameworks.
AI/ML Operations and Model Management
The focus on operationalizing AI/ML capabilities involves efficiently managing and deploying models and data.
EPC AI/ML Operations
- Monitoring, retraining and deploying machine learning models at scale.
- Enhancing the use of open-source tools while ensuring security.
- Promoting the reuse of production models to improve cost efficiency.
COBRRA a.k.a. Continuous Operational Behavior Risk and Resilience Assurance
- Real-time monitoring of model performance in active deployments.
- Identifying data drift to mitigate potential prediction inaccuracies.
- Providing lightweight software for edge environments to enable adaptive learning capabilities in AI systems.
- Implementing live outlier detection for operational scenarios.
Join CDAO Chief Information Officer Daniel Holtzman and Deputy Director for the CDAO’s Advanced Command and Control Accelerator Lindsey Sheppard at the 2025 AI Summit on March 20! Register here.
Advancements in AI Testing Tools and Standards
CDAO integrates cutting-edge tools and standards to ensure reliable and secure AI implementation across the DOD.
JATIC (Joint Artificial Intelligence Test and Evaluation Center)
- Development of enterprise-level TEVV tools for model testing and evaluation.
- Advancing adversarial AI techniques to strengthen AI systems’ robustness, building upon DARPA’s GARD program.
- Creation of AI/ML model cards for transparency about model information and dataset usage.
- Establishing data repositories and ontologies specific to DOD applications.
T&E Infrastructure Gap Studies
- Assessing current and future needs for AI T&E infrastructure, identifying gaps and proposing investments to close these gaps.
- Conducting AI cybersecurity posture studies to evaluate risks and recommend improvements.
Perceptor and Smart Sensor Initiatives
Perceptor
- Core software lifecycle management and user interface enhancements.
- Deployment support for mission partners, including analytic containerization and repository integration.
- Building customer engagement to provide feedback channels and expand analytic-sharing possibilities.
Smart Sensor Program
- Focusing on sensor autonomy and non-GPS navigation technologies.
- Enabling AI-driven reasoning for automated target recognition and data fusion.
- Advancing AI-enabled control platforms for autonomous operations.
Responsible AI Implementation
The CDAO prioritizes responsible AI (RAI) practices to mitigate risk and enhance governance.
RAI Tools and Best Practices
- Creation of RAI tools for governance, auditing and simulation, including an executive dashboard and bias detection systems.
- Developing a Risk Management Framework for evaluating high-consequence AI risks.
- Partnering with academic institutions and building mission partnerships to refine RAI integration.
Goals for FY 2025 AI Budget Allocation
Aligned with its 2025 objectives, the CDAO aims to reinforce the DOD’s competitive edge through the implementation of advanced AI solutions. Primary goals include improving efficiency, risk mitigation and operational excellence in AI systems by focusing on the areas above.
Target Audience Impact
These investments are critical for ensuring defense contractors, the industrial base and technology providers are aligned with DOD’s future operational goals. The 2025 CDAO budget not only signals opportunities for private sector collaboration but also highlights the importance of innovative, responsible and mission-oriented AI solutions for national defense.
By emphasizing comprehensive T&E frameworks, leveraging advanced AI/ML models and promoting responsible AI practices, the DOD continues to forge pathways for effective, efficient and secure applications of enterprise-grade AI technologies.
Join your most beloved peers and fiercest competitors at the year’s definitive AI GovCon conference — the 2025 Artificial Intelligence Summit, from Potomac Officers Club. It’s all happening on March 20 at the Hilton McLean in Virginia. Secure your spot before it’s too late.
