DEPAR TMENT OF DEFENSE

Data, Analytics, and Artificial

Intelligence Adoption Strategy
Accelerating Decision Advantage

Cleared for open publication ▪ June 27, 2023, ▪ Department of Defense ▪ Office of Prepublication and Security Review
 TABLE OF
CONTENTS

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY

 FOREWORD 2

STRATEGIC ENVIRONMENT 3

KEY OUTCOMES 5

STRATEGIC GOALS 7

IMPLEMENTATION 14

CONCLUSION 17

APPENDIX A: QUALITY DATA 19

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY

 “America’s DNA is to innovate ...
and it has repeatedly enabled us to
drive and master the future
character of warfare.”
Kathleen H. Hicks
Deputy Secretary of Defense

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 1

 FOREWORD
The Department of Defense (DoD) has been investing in artificial intelligence (AI) and responsibly fielding
data- and AI-enabled systems for over 60 years. Today, data, analytics, and AI technologies are
increasingly available to DoD Components and providing value to our service members.

Alongside industry’s advancements, DoD has for years made steady and swift improvements to its data
foundation and analytics capabilities: experimenting with AI through research and development,
integrating these technologies into business and warfighting functions, and laying the foundation for
their use at scale. As our investment, experimentation, and innovation continues and accelerates, our
task now is to drive the diffusion of these technologies across the enterprise.

Although our strategic competitors have ambitious aims for AI, the United States and its military possess
strong structural advantages in talent, warfighting experience, technology availability, and systems
integration — not to mention the values that guide everything we do. Equipping our people with the
tools and resources to make better decisions faster will increase the efficiency of DoD business
operations, make our warfighting capabilities and the people who command them more effective, and
create opportunities to employ novel operational concepts.

Responsibly and rapidly realizing the full promise of data, analytics, and AI is not the sole job of a single
organization or program; it’s on all of us. Providing DoD data as an enterprise resource, for instance,
requires more sharing and collaboration, not less. We seek an agile strategic approach that guides
decentralized action across DoD, inspires campaigns of learning, and leverages all our people, processes,
and enabling technologies.

As we have integrated analytics and AI applications, we have observed their benefits and learned crucial
lessons about their limitations. From the boardroom to the battlefield, more work remains, such as
improving data quality and network infrastructure. This Strategy serves as a guide for how we will
strengthen the organizational environment in which DoD deploys data, analytics, and AI capabilities for
enduring decision advantage.

Successfully defending the nation depends on our people. As we have always done, DoD will continue to
trust, support, empower, and invest in our people. We will not outpace our adversaries through
imitation. We will succeed by leading with our strengths: our democratic values, our diverse and open
society, our culture of ingenuity, our second-to-none innovation base, and our globe-spanning network
of Allies and partners. Together, we will harness data, analytics, and AI for the defense, security, and
prosperity of the American people and the world.

Kathleen H. Hicks
Deputy Secretary of Defense

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 2

 STRATEGIC ENVIRONMENT
As the 2022 National Defense Strategy (NDS) makes clear, the United States possesses strengths that our
competitors cannot match, among them our diverse and open society, our culture of ingenuity, our innovation
base, and our globe-spanning network of Allies and partners. The Department leverages these strengths by
distributing authority, empowering leaders in our All-Volunteer Force to innovate at the edge and apply their own
judgment to combine old and new capabilities into superior operational concepts. The latest advancements in
data, analytics, and artificial intelligence (AI) technologies enable leaders to make better decisions faster, from the
boardroom to the battlefield. Therefore, accelerating the adoption of these technologies presents an
unprecedented opportunity to equip leaders at all levels of the Department with the data they need, and
harness the full potential of the decision-making power of our people.

The NDS also describes the need for the United States to sustain and strengthen deterrence against the People’s
Republic of China and other strategic competitors, which have widely communicated their intentions to field AI for
military advantage. Accelerating adoption of data, analytics, and AI technologies will enable enduring decision
advantage, allowing DoD leaders to prioritize investments to strengthen deterrence; link cross-cutting campaign
outcomes that counter our competitors’ coercive measures; and deploy continuous advancements in
technological capabilities to creatively address complex national security challenges in this decisive decade.

The urgency of the strategic environment and the scale at which the Department must operate are formidable.
The Department is well-positioned to excel because it has established a foundation of strategic guidance informed
by lessons learned from hands-on initiatives over the last several years.1 The Department’s first AI Strategy,
published in 2018, and revised Data Strategy, published in 2020, are two of these foundational efforts. The 2018
AI Strategy emphasized the need to build centralized infrastructure for AI development, to bridge AI technology
developments from the Department’s research and engineering communities, and to exercise international
leadership in military ethics and AI safety. The 2020 Data Strategy envisioned the Department as a data-centric
organization that can employ data supporting advanced capabilities for operational advantage and increased
efficiency, and oriented enterprise data management activities toward the VAULTIS goal framework.2

1. This guidance includes the DoD AI Strategy (2018), the DoD

Digital Modernization Strategy (2019), the DoD Data
Strategy (2020), the DoD Enterprise DevSecOps Strategy
Guide (2021), the DoD Software Modernization Strategy
(2022), the Trusted AI and Autonomy Critical Technology
Roadmap (2022), and the DoD Zero Trust Strategy (2022).

2. The 2020 DoD Data Strategy outlined the following seven

goals (VAULTIS): Visible – Consumers can locate the needed
data. Accessible – Consumers can retrieve the data.
Understandable – Consumers can find descriptions of data
to recognize the content, context, and applicability. Linked –
Consumers can exploit complementary data elements
through innate relationships. Trustworthy – Consumers can
be confident in all aspects of data for decision-making.
Interoperable – Consumers and producers have a common
representation and comprehension of data. Secure –
Consumers know that data is protected from unauthorized
use and manipulation.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 3

 Since these strategies were published, industry has produced more tools, platforms, and services for federated
environments, enabling more effective, decentralized data management, and analytics and AI development.
Adoption of these commercial offerings has allowed organizations within the Department to focus on
necessary internal transformation efforts and deploy government-owned tools, services, and platforms for
military use cases. The Department has matured collaboration with academia, industry, as well as Allies and
partners, and promoted best practices on data management, responsible AI, and AI readiness.
Experimentation and fielding have resulted in a deeper understanding of the degrees of data quality and
availability required to develop and deploy advanced analytics and AI capabilities at scale.

This DoD Data, Analytics, and AI Adoption Strategy builds upon and supersedes the 2018 AI Strategy and the
2020 Data Strategy to continue the Department’s digital transformation. The Department will continuously
seize opportunities presented by iterative technology advancements, at the speed of relevance and at the
scale of our global mission. To do so, the Department requires a unified approach across data, analytics, and AI
activities; an educated, empowered workforce skilled at incorporating commercial teams and tools; continued
advanced research and rapid experimentation; and effective integration with our Allies and partners. The
Department cannot succeed alone. Our integration of data, analytics, and AI technologies is nested within
broader U.S. government policy, the network of private sector and academic partners that promote
innovation, and a global ecosystem. We need a systematic, agile approach to data, analytics, and AI adoption
that is repeatable by all DoD Components. This strategy outlines our approach to improving the organizational
environment within which our people can deploy data, analytics, and AI capabilities for enduring decision
advantage.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 4

 KEY OUTCOMES
As a result of implementing this strategy, DoD leaders and warfighters will be able to make rapid,
well-informed decisions by expertly leveraging high-quality data, advanced analytics, and AI as part
of a continuous, outcome-driven, and user-focused development, deployment, and feedback cycle.

The Department’s investments in data, analytics, and AI will address key operational problems identified in the
2022 NDS, fill validated gaps to enhance the warfighting capabilities of the Joint Force, and strengthen the
enterprise foundation required to sustain enduring advantages. Fielding data, analytics, and AI capabilities across
this continuum from the boardroom to the battlefield recognizes that warfighting decision advantage is enabled
by hundreds, or thousands, of decisions made by personnel and program offices at great distances from the
frontline. Strengthening decision advantage for the Department’s warfighting and business operations is key to
maintaining a resilient future force that can address a broader array of operational problems, dynamically
campaign and deter, and prevail in conflict, if necessary.

Decision advantage is a competitive condition characterized by the following outcomes:

• Battlespace awareness and understanding

• Adaptive force planning and application
• Fast, precise, and resilient kill chains
• Resilient sustainment support
• Efficient enterprise business operations

Agile, user-focused, product-centric development is essential to achieving these outcomes because humans and
machines will work together in the responsible, effective employment of data, analytics, and AI-enabled
capabilities.

Today, there are multi-disciplinary teams throughout the Department that leverage common technology
development best practices. These practices include:

• Employing Agile development fundamental principles and approaches

• Building intuitive interfaces to accelerate human adoption of new technology
• Developing products with cross-functional teams focused on customer needs
• Offering product portfolios with shared digital foundations
• Experimenting with minimum viable products in operational environments to identify new concepts for use,
improve capability, and manage emergent risks

More is needed now, and at scale. The Department will pursue a multi-disciplinary approach and implement
these best practices to strengthen its technology, human capital, processes, and culture. This approach has
implications analogous to pivoting from a heavy armor force to one with greater maneuverability. The
Department will enhance its competitive edge through a vigorous and continuous capability delivery pipeline that
can respond with agility to changing environments and technologies.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 5

 The Department’s agile approach to adoption (Figure 1) ensures a tight feedback loop between
technology developers and users through a continuous cycle of iteration, innovation, and
improvement of solutions that enable decision advantage. Practicing agility and learning by doing will
accelerate deployment speed–measured in hours or days, not months or years. Creating effective,
iterative feedback loops among developers, users, subject matter experts, and test and evaluation
(T&E) experts will ensure capabilities are more stable, secure, ethical, and trustworthy.

Figure 1: Employing an Agile Approach to Adoption to Scale Decision Advantage Outcomes

An agile approach to adoption emphasizes speed of delivery and continuous improvement, prioritizing
outcomes over processes. Valuing speed necessitates organizational agility and learning through early
and ongoing real-world feedback. The Department will move toward greater integration,
transparency, and knowledge sharing across organizational boundaries. Increased diffusion of data,
analytics, and AI technologies will introduce technical vulnerabilities that require rigorous protection
measures. These risks will be managed not by flawless forecasting, but by continuous deployment
powered by campaigns of learning. Developing capability in this way enables responsibility, ensuring
not only the sustained quality, stability, and security of DoD systems, but also providing the means by
which engineers can reduce unintended bias and instill justified confidence with their users.

The Department’s agile approach to adoption ensures a tight feedback loop between
technology developers and users through a continuous cycle of iteration, innovation, and
improvement of solutions that enable decision advantage.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 6

 STRATEGIC GOALS
The Department will focus strategic efforts on several interdependent goals that support the DoD AI Hierarchy of
Needs (Figure 2). The AI Hierarchy of Needs is a pyramid with quality data as its foundation since all analytic and AI
capabilities require trusted, high-quality data to support decision makers. The next layer in the Hierarchy is
insightful analytics and metrics, the foundational models and visualizations required for DoD leaders to understand
their domain and the key variables impacting outcomes in those domains. At the top of the pyramid is Responsible
AI, the Department’s dynamic approach to the design, development, deployment, and use of AI capabilities in
accordance with the DoD AI Ethical Principles while delivering better, faster insights and improved mission
outcomes.3 The layers of the Hierarchy are supported by robust sets of processes. Increased data quality and
insightful analytics are achievable through effective enterprise data governance. Sound assurance processes for
testing, evaluation, validation, and verification are imperative for Responsible AI. Around the pyramid are enablers,
such as digital talent management, that help sustain the Hierarchy of Needs.

Figure 2: Strategic Goals and the AI Hierarchy of Needs

The Hierarchy is helpful as a framework for assessing DoD AI readiness, and for guiding the Department’s goals to
accelerate adoption of data, analytics, and AI technologies to build enduring decision advantage. These
interdependent goals, and their supporting activities and investments, cut across technology, human capital,
process, and culture areas, and are described in further detail on the next few pages.

3. For more information on the Department’s Responsible AI plan, see the “US Department of Defense Responsible Artificial
Intelligence Strategy and Implementation Pathway.”

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 7

 IMPROVE FOUNDATIONAL DATA MANAGEMENT: Increase the quality and availability of relevant
DoD data to support advanced analytics and artificial intelligence capabilities.

Consistent with the Deputy Secretary of Defense’s Memorandum

on Creating Data Advantage, the Department will value data as a
product, ensuring the responsible collection, storage, and
management of relevant data to support enterprise needs. All
DoD data is an enterprise resource. The Department will adapt
and implement open standard architectures while abiding by
existing DoD cybersecurity policies and heed industry best
practices for data ethics, data protection, and design as
technology evolves. The Department will also continue to make
its data more visible, accessible, understandable, linked,
trustworthy, interoperable, and secure (VAULTIS). Components
will assess their data across lifecycles using the data quality
dimensions and the VAULTIS framework first outlined in the 2020
DoD Data Strategy.4

The Department’s data management focus will initially prioritize

improving data quality and managing data as a product in support
of the Secretary of Defense's priority areas. To improve DoD data
quality across the enterprise, the Department will develop and
implement a decentralized network among data providers and
users. This network will consist of both process-based and
technical components, distributing ownership across data
domains and treating data as a product. Instead of designating a
centralized data team responsible for managing all data across the
enterprise, data domain owners and data product teams will be
responsible for managing the data products they own and
produce.

Data products will be designed, built, and maintained with the

needs and requirements of its users in mind, just like a traditional
product. By treating data as a product, DoD Components can
stimulate a culture of data sharing and reuse under appropriate
circumstances, which breaks down data silos and promotes cross-
functional collaboration. This product orientation ensures that
data is properly managed and governed, with clear accountability,
quality and interface standards, and access controls. While this
data management approach will not reduce DoD’s organizational
complexity, over time it will improve operational and analytical
data quality, reduce data backlogs, lower data storage costs, and
reduce data redundancy. These improvements will allow DoD
organizations to better leverage their data products and make
more effective, data-driven decision-making.

4. See Appendix A for further detail.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 8

 DELIVER CAPABILITIES FOR ENTERPRISE BUSINESS AND JOINT WARFIGHTING IMPACT: Enhance
and/or generate business analytics and warfighting capabilities with data, analytics, and AI
technologies for improved decision advantage outcomes

An integrated, agile approach with a strong focus on data quality will ensure the Department fields responsible
solutions that most appropriately address organizational needs. There is an urgent need to surge digital support to
address Joint capability gaps at the operational to strategic levels in direct support of operational commands. Building
momentum toward transformation requires demonstrating tangible, near-term results. For example, a user
requesting an advanced decision support model with recurring excursion capability could field a simpler analytics
dashboard in the near-term to better understand current state and shape more sophisticated questions. The
Department will continue to focus on advancing business analytics, to continuously improve data quality through use,
and to provide the more complete picture DoD personnel need to make better reactive and proactive decisions.

The Department will design and test analytics and AI-enabled solutions side-by-side with stakeholders, business and
warfighter, and demonstrate capabilities via robust campaigns of learning to account for different operational
environments. Some users may have access to abundant data and high-speed processing while other users face
limited bandwidth at the tactical edge. Designers will identify these constraints early in the development process.
The Department will conduct continuous experiments, and integrated multi-lateral exercises with high-impact Joint
Force use cases that can inform additional investments in interoperable enterprise-level, warfighting capabilities.
Components will advocate for investments in data, analytics, and AI that will generate these capabilities. Capabilities
deemed successful after rigorous testing will receive clear pathways to sustainment and wide-scale adoption.

STRENGTHEN GOVERNANCE AND REMOVE POLICY BARRIERS: Ensure responsible behavior,

processes, and outcomes while accelerating the pace of adoption for data, analytics, and AI
technologies across the Department

The Department’s approach to integrated data, analytics, and AI governance will account for the scale of the
organization, its distributed authority structure, and high degrees of variance in data maturity among Components.
Further, the Department’s information enterprise is vast and global, with strong dependencies among different IT
systems, support teams, programs, and activities. For example, a key pillar of the 2022 DoD Zero Trust Strategy hinges
upon the establishment of effective enterprise data governance, and the fielding of analytics and AI to secure our
networks, applications assets, and services. Similarly, records management processes and procedures are another
dependency. The 2023 DoD Records Strategy makes clear that records are data and will be curated, automated, and
governed. Treating records in this way will increase trust in data and analytics products.

These examples of dependencies underscore the complexity of the Department’s policy environment and the
necessity of consensus building. The DoD data, analytics, and AI leadership community will build consensus around
responsible development practices that enable mission owners and other governance bodies while serving as a
demanding customer of enabling capabilities. Data, analytics, and AI governance will be risk-adjusted, streamlined,
and data-driven, and focused on collaborative learning. Generating consensus and collaboration is critical to
mitigating known and emergent policy barriers and aligning technical interfaces to hasten responsible adoption.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 9

 In support of a collaborative approach to enterprise governance, DoD Components will identify clear leaders for data-
related transformation, strengthening accountability across the Department. Enterprise-level governance initiatives
will prioritize key areas including data management, cybersecurity, requirements, joint interoperability, records
management, and Responsible AI. The success of these areas requires Components to continue to follow responsible
security procedures, review current policies and processes on data aggregation and classification, revise issuances as
needed, and comply with updated guidance. The Department will approach reforms with speed and responsibility:
targeting identified policy revisions to improve agility, speed of capability deployment, and scalability; while upholding
a steadfast commitment to lawful and ethical behavior; and protecting privacy and civil liberties. Components’ efforts
to implement this Strategy will be supported by governance policies, processes, and structures that are mission-
driven, responsive to need, minimally prescriptive, and based on best practices. The Department will reduce
institutional barriers, including those that unnecessarily inhibit collective research and development, planning,
interoperability, intelligence, and information sharing. The Department will lead across the U.S. government to
inform technology and information release processes, expand release authorizations, and redefine dissemination
controls to facilitate information exchange for mutual benefit.

INVEST IN INTEROPERABLE, FEDERATED INFRASTRUCTURE: Optimize the Department’s

federated infrastructure to support scaling data, analytics, and AI adoption and improve
interoperability.

The Department will invest in

abundant, flexible, secure, and
jointly interoperable
infrastructure that is scalable for
the needs of users. Data,
analytics, and AI capability
development requires
tremendous computing power
and demand will grow
exponentially as adoption scales.
When appropriate, this
infrastructure will be automated,
including measures to implement
DoD technology policies (e.g.,
Continuous Authorization to
Operate), its status reporting,
and, critically, user access to
Figure 3: Balancing Tradeoffs with Shared Services
mission- relevant data, analytics,
and AI platforms. Though government-led and designed, DoD infrastructure will adopt open standard architectures for
industry and trusted partners to facilitate collaborative and continuous experimentation. As described in the DoD
Software Modernization Strategy, the Department is an “enterprise of enterprises.” Thus, the Department’s
infrastructure supporting data, analytics, and AI technologies will remain federated. The Department will continue to
centralize some decisions and services while others remain decentralized to address unique mission needs. To strike
the optimal balance of platforms and services, the Department will assess infrastructure based on outcome
commonality and implementation complexity (depicted in Figure 3).

13

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 10

 While it is important to strive for enterprise scaling economies, leaders must also understand the extent to which
their Component's data, analytics, and AI requirements are truly specialized. Organizations with common digital
service requirements and joint use cases, particularly when migration would be relatively simple, will move quickly to
adopt shared services to meet their needs. Organizations’ infrastructure must remain decentralized if they have
highly specialized requirements or if leaders determine that the costs outweigh the benefits of consolidating within a
shared platform or suite of services. One notable exception to this logic is AI-specific infrastructure. The scarcity of
expertise required to customize safe and reliable AI-enabled systems remains a limiting factor. Thus, DoD
Components will seek to centralize certain AI development platforms for continuous experimentation and
deployment, and then maximize the adoption of responsible AI shared services.

ADVANCE THE DATA, ANALYTICS, AND AI ECOSYSTEM: Strengthen intergovernmental,

academic, industry, and international partnerships to enable adoption of data, analytics, and AI
technology.

The Department will advance progress toward a

robust national and international ecosystem
that facilitates improved intergovernmental,
academic, industry, and international
collaboration on data, analytics, and AI
technology. The Department cannot succeed
alone. Through domestic and international
engagements, the Department will collaborate
on common challenges, further shared
interests, promote democratic norms and
values, and increase interoperability with
partners. The Department will cooperate with
Allies and partners to leverage comparative
advantages and allow for interoperability in
tactics, institutions, and strategies related to
data, analytics, and AI. Where appropriate, the
Department will continue exporting key
technologies and sharing data to ensure our
Allies and partners remain agile and capable of
rapidly employing advanced analytics and AI
innovations.

To obtain and integrate proven solutions in

collaboration with industry, the Department
will follow an “adopt-buy-create” framework
aligned with the DoD Software Modernization
Strategy and Office of Management and Budget
(OMB) Circular A-130, Managing Information as
a Strategic Resource.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 11

 DoD leaders will first seek to adopt solutions that are already Joint- or Component-sponsored before exploring
capabilities available on the open market. When DoD-owned shared services are unavailable, the Department
will challenge vendors to solve specific business and mission problems, while designing acquisition strategies
to avoid vendor lock-in. DoD customers with clean, high-quality data can seek commercially available analytics
and AI capabilities while retaining appropriate data rights. Government contracts for commercial solutions will
ensure the Department’s capability pipelines address evolving requirements while balancing protection of
industry intellectual property. Commercial solutions may not meet all mission requirements, but they can
provide best-in-class capabilities for many dual-use applications. It is often in the Department’s interest to
procure software and support for these commercial solutions, freeing up DoD engineers for inherently
governmental challenges. Fielding web-based, cloud-based, and/or Application Programming Interface (API)-
first applications create more opportunities for rapid, enterprise scalability; continuous integration and
delivery; and increased economies of scale. APIs also allow for a more open exchange with diverse data
sources, regardless of origin, giving DoD leaders and warfighters greater access to the information they need
to make more timely and accurate decisions.

The Department will only create solutions when those applications aid DoD-specific missions and cannot be
readily adopted from commercial or existing solutions. These Defense-specific applications often involve real-
time, mission-critical, embedded software coupled to customized DoD hardware. Creating an ecosystem that
fosters competition and collaboration is essential for the development and deployment of AI-enabled systems,
particularly next generation capabilities for use in Joint warfighting. When supported with open standard
architectures, the Department’s data, analytics, and AI ecosystem will facilitate commercial competition based
on model performance and promote collaboration with trusted international and intergovernmental partners.
Commercial models verified, validated, accredited, and fielded through this process can then be customized
for DoD use.

Program managers will consider acquisition strategies that leverage international partnerships and
supportability planning to improve economies of scale, strengthen the defense industrial base, and enhance
Ally and partner capabilities. Further, the Department will embrace initiatives aimed at creating standard
language necessary for data, analytics, and AI technology contract problem statements and agreements.
When applicable, the DoD will leverage resources such as the federated data and model catalog and AI-
enabled enterprise tools to expedite acquisitions. The Department will mature and expand acquisitions
policies that increase government visibility of the ownership, labeling, maintenance, and classification of data.
The Department has made progress on acquisition innovation and will continue to encourage personnel to
embrace risk and learn by doing. Contracting must be as agile as the industry with which it is engaging. The
Department will continue to adapt best practices from nontraditional partners to ensure cutting edge
solutions are delivered responsibly at the speed of relevance.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 12

 EXPAND DIGITAL TALENT MANAGEMENT: increase hiring, training, and retention for the most
critical data, analytics, and AI-related work roles.

The Department’s ability to oversee and adopt data, analytics, and AI capabilities depends on the strength of its
workforce, and continued growth in technical skills. The Department will ensure it is identifying and employing the
talent it already possesses, then focus resources on those initiatives and tools that best attract, recruit, train, and
retain a truly innovative workforce across digital talent work roles. The Department will also identify and train
nontechnical personnel who will lead and oversee a culture of innovation that advances the responsible use and
adoption of data, analytics, and AI capabilities.

The Department’s infrastructure and digital ecosystem remain obstacles to hiring talent from the cutting edge of
the private sector. While the Department takes all necessary steps to enhance its technical foundation, talent
management efforts and resources will focus on upskilling and reskilling Service members and civilians in the work
roles that DoD Components have determined are the most important for addressing the needs of the
Department. For instance, Service members and civilians with domain knowledge and basic digital skills will be
upskilled or reskilled with targeted training and hands-on opportunities to serve in work roles such as data
architect, data steward, and user experience designer within the newly expanding Department-level workforce
framework.

Components will draw personnel identified for upskilling and reskilling opportunities from the Department’s Total
Force. Active and Reserve Service members often possess high-demand digital skills not directly aligned to their
occupational specialty. To retain talent, the Department will create more flexible service structures that cultivate
and reward strengths and avoid penalizing personnel for selecting non-traditional career paths.

While the DoD digital workforce

grows and strengthens through
upskilling and reskilling, the
Department will also think
differently about maturing
pipelines to attract, recruit, and
flexibly hire digital talent.
Components will use existing hiring
authorities and retention tools at
their disposal, and institute
reforms where appropriate to
allow for maximum flexibility in
garnering digital talent.
Additionally, the Department will
execute a series of pilots to
identify organic talent, validate
barriers and blockers presented,
and establish a cadre of Service
members and civilians from across
the DoD enterprise to build and
apply digital solutions for the most
difficult missions.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 13

 IMPLEMENTATION

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 14

 IMPLEMENTATION

The Chief Digital and Artificial Intelligence Office (CDAO) will lead and oversee the implementation of this Strategy.
The CDAO will collaborate on implementation with Components through the CDAO Council, the senior leadership
body that governs and coordinates the Department’s integrated data, analytics, and AI enterprise.5 For certain
issues, the CDAO Council recommends decisions on data, analytics, and AI to the Deputy’s Management Action
Group, the Deputy's Innovation Steering Group, and the Deputy’s Workforce Council. The CDAO will conduct an
annual review of the Strategy and report results through the CDAO Council. The CDAO Council and supporting
forums will ensure the exchange of challenges, lessons, and best practices gleaned from across the Department, and
oversee initiatives for digital talent development, leadership, and culture as critical enablers.

IMPLEMENTATION PLANNING GUIDANCE

DoD Components have assigned oversight for data, analytics,

and AI differently based on their missions, governing laws and
applicable procedures, and organizational structures. The state
of data maturity across the Department varies significantly from
Component to Component. Thus, Components will tailor
implementation according to self-assessed data maturity levels,
mission parameters, and pertinent laws. This strategy does not
prescribe that Components reorganize in the same fashion as
the DoD CDAO. However, within 60 days of publication,
Components will designate their team or office of primary
responsibility for implementing this strategy and identify any
teams or offices with contingent responsibilities. After analyzing
the goals of this strategy, Component leaders may determine
that multiple teams will be responsible for implementation.

Based on lessons learned from implementing the previous DoD

AI and Data Strategies, outcomes-based performance indicators
will be established, refined, and monitored in coordination with
the CDAO Council as an integrated part of the Department’s
enterprise performance analytics framework. To further aid
Component-level decision-making and execution, the CDAO will
publish expanded implementation guidance and additional
appendices to this strategy. The expanded implementation
guidance will outline the process by which the CDAO and CDAO
Council collaborate with the Components to create agreed upon
strategic performance measures linked to this Strategy’s key
outcomes and strategic goals. This collaborative process will
ensure measures are supported by authoritative data sources
and maximize the use of automated data collection methods for
efficient performance monitoring. Where Authoritative Data
Sources do not exist, the CDAO will assist in their construction
and maturation until these sources meet senior leader decision
requirements.

5. Standing membership for the Council is at the Deputy and/or Vice level (or their delegated representatives) of the following organizations: the Offices of
the Under Secretary of Defense for Research & Engineering, Acquisition & Sustainment, Policy, Comptroller, Personnel & Readiness, and Intelligence &
Security; the Office of Cost Assessment and Program Evaluation; the Office of the Chief Information Officer; the Joint Chiefs of Staff J3 and J6; the
Combatant Commands; the Military Departments; and the National Guard Bureau.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 15

 IMPLEMENTATION

RISK MANAGEMENT

Data, analytics, and AI-enabled technologies are having profound impacts on the ways we work, live, and engage
with one another. Aided by training and user-friendly interfaces, the integration of these capabilities within the
Department’s systems and activities will continue to increase until they are nearly ubiquitous. The anticipated
breadth of adoption by the Department and its competitors, and the nonlinear evolution of these technologies,
present strategists with unique foresight challenges, especially for resource planning. For data, analytics, and AI
technologies, Components will identify and adopt resourcing schemes, processes, and assessment tools that offer
leaders maximum flexibility to compliantly deliver iteratively developed capabilities at speed. Each iteration or
release cycle will incrementally reduce risk while rapid feedback loops ensure that Components more consistently
meet user demands. By staying adaptive, the Department can maintain a deep understanding of the specific needs
of end users and adjust more quickly to changes in the security environment.

This strategy’s learning-based, agile approach to adoption and emphasis on data quality also mitigate
implementation risks. Though strategic success depends upon a high degree of decentralized execution, the CDAO
Council and Component leaders can learn from one another and coordinate to manage critical dependencies
among the goals in this strategy and parallel goals in other DoD strategies. As the warfighting use cases for
analytics and AI technologies continue to expand, we can expect our strategic competitors to field them to enhance
their capabilities. Our adversaries will also continue to target U.S. technologies for theft and exploitation.
Therefore, the Department will employ development approaches that allow us to move quickly, protect our
advantages, and abide by our laws, policies, and values. The Department recognizes the privacy and civil liberties
challenges posed by data, analytics, and AI capabilities and will establish transparent governance and compliance
processes that address the full scope of these potential risks. By focusing on data governance and data quality for
analytics and AI development, the Department can mitigate certain risks, including the replication of unintended
bias across the enterprise. Component leaders and technologists remain committed to the objectives of the DoD
Responsible AI Strategy & Implementation Pathway and to developing AI capabilities that are responsible,
equitable, traceable, reliable, and governable.

19

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 16

 CONCLUSION

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 17

 The Department’s scale and warfighting mission are unique, but organizations from all sectors have overcome
similar challenges and have harnessed the benefits of digital transformation. DoD leaders will have access to high
quality data, advanced analytics, and AI capabilities to make timely and well-informed decisions to defend the
homeland, deter aggression, and win in conflict. Our military competitors are integrating these same technologies
for their own advantage. We cannot afford to wait and, moreover, we cannot succeed alone. This strategy’s
approach embraces the need for speed, agility, learning, and responsibility. Pursuing this agile approach and
focusing activities on the goals outlined in this strategy will allow the Department to adopt data, analytics, and AI-
enabled capabilities at the pace and scale required to build enduring decision advantage. If we confront our
challenges holistically and refuse to accept the status quo, we will accelerate data, analytics, and AI adoption and
continuously deploy creative solutions for the defense, security, and prosperity of the American people.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 18

 APPENDIX A: QUALITY DATA
This appendix provides additional guidelines for
improving the Department’s foundational data
management to increase the quality and availability of
relevant DoD data. User needs and local requirements
are the basis of data quality. Therefore, Components
and subordinate organizations will develop data quality
implementation guidance and associated metrics.
Since this appendix retains the VAULTIS framework,
existing Component-level implementation plans based
on the 2020 Data Strategy are still sound. However,
over time, Components will evolve their data
management plans to align more closely with this
Strategy.

As Components develop and implement their data

quality plans, data sets targeted for improvement will
be founded on metadata to allow for data search and
discovery; prioritized for relevance and mission value;
acted upon such that improvement efforts are
appropriately resourced; monitored to measure and
report quality levels on priority data sets across their
lifecycles; and corrected through cyclical processes to
continuously address quality degradation.

Two tools for use in data management planning are the

data quality dimensions and the VAULTIS framework.
These tools represent interrelated and mutually
supporting concepts that apply across data’s lifecycle,
from creation to disposal. Poor quality data will
inevitably undermine data trustworthiness, raise
security concerns, and negate the utility of the VAULTIS
framework. Conversely, even high-quality data that
does not align to the VAULTIS framework will be of
limited value to analytical and AI efforts.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 19

 APPENDIX A: QUALITY DATA

1. DATA QUALITY DIMENSIONS

Data owners may categorize data sets at different levels of quality over time because of age, prioritization, initial
condition, and other factors. Therefore, data quality dimensions are relative, and owners will assess the
dimensions across the data’s lifecycle. The table below provides sample assessment questions for each dimension.

Dimension Assessment Questions

Accuracy: Data that correctly reflect proven,  How frequently do values not align to their assigned
true values or the specified action, person, or format?
entity. Accuracy includes data structure,  How frequently do data values match ground truth?
content, and variability.  How is error measured? Is it tolerable for the specified
purpose?
Completeness: The data present at a specified  Is there known data that would make the set more
time contain the expected information or complete?
statistics, as measured at the data set, row, or  Does the data set contain sufficient breadth of
column level. information to contextualize the data for its purpose?
 What fields in the data expect some null values? How
often are null values present?
Conformity: Data sets follow agreed upon  Does the data’s format match the applicable
internal policies, standards, procedures, and standard(s)?
architectural requirements.  Is the data set architecture published and available?
 Are there database constraints implemented in
accordance with internal policies, standards, and
procedures to prevent erroneous input?
Consistency: The degree to which a value is  Are there other data sets that reference values in this
uniformly represented within and across data data set?
sets.  Are there discrepancies?
Uniqueness: Ensures there is a one-to-one  Are there other Authoritative Data Sources that serve
alignment between each observed event and the same function?
the record that describes such an event.  Are there duplicate records in this data set?
Integrity: A data set’s pedigree, provenance,  Are there opportunities for data to be tampered with,
and lineage are known and aligned with misreported, degraded, corrupted, poisoned, or
relevant business rules. otherwise altered during the collection, storage,
processing, or transmission processes?
 How often are data quality checks conducted to address
poor data quality?
 Does the data cleaning process result in data that can be
trusted?
Timeliness: Measures the time between an  How frequently do supported data consumers require
event occurring and the data’s availability for updates?
use.  Does the data purpose require reduced latency?

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 20

 APPENDIX A: QUALITY DATA

2. VAULTIS FRAMEWORK

The VAULTIS framework discussed in the 2020 DoD Data Strategy remains a useful tool for Components to organize
and monitor their data management plans. The key attributes and associated assessment questions include:

Attribute Assessment Questions

Visible: Consumers can locate  Are authorized users able to discover the existence of data that is of
the needed data. particular interest or value?
 Are data stewards, data custodians, and functional data managers
assigned responsibility for identifying, registering, and exposing data to
make it easily discoverable across the enterprise, and to appropriate
external partners?
 Are users able to discover and rapidly identify who is responsible for
specific data products, the location of data products, the types of data
products available, and the means of accessing the data products?
Accessible: Consumers can  Are authorized users able to obtain the data they need when they need
retrieve the data. it?
 Are data, including warfighting, intelligence, and business data,
accessible to authorized users?
Understandable: Consumers can  Are tools in place that allow users to securely aggregate, compare, and
find descriptions of data to analyze the data?
recognize the content, context,  Are the data labeled and formatted appropriately to support large scale
and applicability. analytics?
 Does the metadata associated with the data contain the contextual
relationships and business rules?
Linked: Consumers can exploit  Are the data linked such that relationships and dependencies can be
complementary data elements uncovered and maintained?
through innate relationships.  Does the organization use industry best practices for open data
standards, data catalogs, and metadata tagging?
 Are the data unnecessarily siloed or do they support connections across
disparate sources?
Trustworthy: Consumers can be  Do the data represent a source of truth?
confident in all aspects of their  Must the data undergo additional vetting to ensure it can support
use of the data for decision- decision-making?
making.  Can users interpret and analyze data without concern for flawed
assumptions, resulting in potentially fatal outcomes?
 Does the data reflect metadata or context necessary to enable
consumers to make appropriate judgments about whether or how to
rely on the data?

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 21

 APPENDIX A: QUALITY DATA

Attribute Assessment Questions

Interoperable: Consumers and  Can the data be exchanged between systems and users while
producers have a common maintaining its quality and usability?
representation and  Do the data have semantic as well as syntactic interoperability based on
comprehension of data. common data formats and machine-to-machine communications?
Secure: Consumers understand  Are the data protected while at rest, in motion, and in use (within
data safeguarding applications, with analytics, etc.)?
responsibilities, follow  Does the organization use a disciplined approach to data protection,
classification management such as attribute-based access control, across the enterprise?
procedures, and know that data  Are protective mechanisms (e.g., security controls) in place for
is protected from unauthorized credentialed users to ensure that access is permitted in accordance
use and manipulation. with applicable laws, regulations, and policies?
 Are classified management procedures in place to ensure that data,
aggregated or not, are safeguarded and in accordance with laws,
regulations, and policies?

Taken together, these guidelines provide a foundation for organizations to greatly enhance the quality of high
priority data sets in effective and efficient ways. Absent an organized approach based on tested criteria,
Components will face significant challenges developing the level of data quality required to address analytics and
artificial intelligence requirements. As such, organizations will use collaborative forums, such as the CDAO Council
and its supporting forums, to flatten communications, enhance visibility, and ensure alignment across approaches
to data quality improvement and maintenance.

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY 22

 26

DOD DATA, ANALYTICS, AND ARTIFICIAL INTELLIGENCE ADOPTION STRATEGY