In early April the White House Office of Management and Budget issued a memorandum encouraging federal agencies to prioritize recruitment and development of technical talent in artificial intelligence (AI). The memorandum, in response to executive order Removing Barriers to American Leadership in Artificial Intelligence, outlined the benefits of prioritizing AI workforce development including increasing capability for responsible AI innovation, providing the federal workforce pathways to AI up-skilling, and assisting employees in applying AI to their work. The memorandum also called on federal agencies to provide “sufficient and periodic training, assessment, and oversight” for staff to act on the AI' s output and manage associated risks.

On a related front, the DoD’s Chief Digital and Artificial Intelligence Office (CDAO) has defined workforce talent and workforce training as critical dimensions along the pathway to AI readiness. At the SEI’s AI Division, we’re creating AI workforce development content to enhance digital transformation initiatives by establishing trust and proficiency among the developers and the users of AI systems. As we’ve seen with our mission partners such as the Office of the Under Secretary of Defense for Research and Engineering (OUSD(R&E)), promoting AI literacy results in better solutions architecture by providing the knowledge necessary to properly develop and maintain systems while increasing user adoption rates. Through numerous discussions with various DoD mission partners around enabling their workforces to implement AI solutions, the SEI AI Workforce Development team has created 10 recommendations that organizations need to know about AI workforce development. Because of the number of individuals within any organization that will be impacted by their commitment to embrace and develop an AI culture, some of these recommendations are designed to develop AI talent among the 11 roles most directly involved with AI and identified by the Defense Cyber Workforce Framework (DCWF):

• AI adoption specialist
• AI innovation leader
• AI risk & ethics specialist
• AI test & evaluation specialist
• AI/ML specialist
• data analyst
• data architect
• data officer
• data operations specialist
• data scientist
• data steward

Other recommendations can be directed to roles less directly involved, such as leadership within the organization, or even the end users of the AI applications.

1. Create immersive, hands-on training to teach AI effectively. This training allows individuals to work directly with AI systems, apply theoretical knowledge, and solve real-world problems. This experiential approach helps build deeper understanding among students by offering the opportunity to experiment with data, develop models, and interpret results. It has also proved to be more effective at retaining new knowledge than lecture alone. In the DoD context, the AI Division’s Workforce Development team builds content to allow students to train like you fight on simulated real-world mission systems.
   
2. Map the DoD workforce to AI-specific roles, such as those defined by the DCWF, to ensure AI training aligns with organizational needs and employee capabilities. By clearly defining AI roles—such as data scientists, AI/ML specialists, application developers, AI-savvy system engineers, business analysts, and non-technical users—organizations can tailor training programs to address both foundational AI skills and advanced expertise. Once roles are defined, the knowledge, skills, abilities, and tasks necessary for those roles also need to be defined. For example, the DCWF maps 71 knowledge, skills, abilities, and tasks (KSATs) to the AI/ML specialist role. However, many of these 71 KSATs also map to one or more of other 11 data/AI DCWF roles and can be tailored to the specific mission of an organization. Aligning training to these KSATs allows some of the developed content to be used for many roles instead of just one, while also being more effective for those roles in that organization. This approach optimizes the distribution of AI knowledge, allowing employees to realize their responsibilities by augmenting their tasks with AI technologies. Time saved reusing KSAT-aligned content can then be better spent on all the developing impactful content for the many unique KSATs that are specific to one of the 11 roles.
   
3. Establish a strategy for consistent AI engineering literacy and approved tools within organizations. Currently, 78 percent of AI users are bringing their own AI tools to their tasks and learning different terminology and knowledge. A consistent baseline knowledge of AI ensures that all employees, regardless of role, can effectively engage with AI technologies. Establishing foundational AI literacy allows organizations to create a common language around AI. This will also result in better collaboration between technical and non-technical teams. Finally, foundational literacy prevents misunderstandings about AI’s capabilities so that decision-makers have realistic expectations and AI initiatives are aligned with mission goals.
   
4. Regularly assess training modules to ensure that they are efficient and effective. Psychometric methods provide the tools necessary to validate assessment—to determine if the assessment measures the training as expected—and to open new possibilities for adaptive learning interventions. Assessment frameworks that support performance-focused evaluations, new measurement techniques that integrate machine learning and AI, and emerging real-time methods that assess practice for in-the-moment trainee profiles present new opportunities to support effective training outcomes and model trainee performance with precision.
   
5. Identify and upskill talent from the current workforce to fill AI roles. AI job postings are growing 3.5 times faster than for all other jobs. Leveraging existing knowledge of the company’s processes, culture, and goals, while addressing the growing demand for AI skills, provides strategic and financial benefits to organizations. Developing internal talent is often more effective than external hiring. Employees already possess domain-specific expertise that can enhance the effectiveness of AI applications in business. Upskilling existing employees also fosters greater engagement and retention. Additionally, identifying which of the 11 DCWF data/AI work roles that an existing employee fits best within can leave them more satisfied with the new position and the organization better prepared for AI systems development.
   
6. Use generative AI to create educational content for AI learning. Generative AI models can rapidly produce tailored learning materials, such as tutorials, quizzes, and videos, which are tailored to meet different learning levels and styles. Incorporating generative AI into the content creation process can also help ensure that learning materials be continually updated with the latest concepts and technologies, providing learners with relevant and timely knowledge. It is important, however, that instructors validate the materials produced by generative AI tools, since there are no assurances of correctness. Generative AI has the potential to improve efficiency by increasing the automation of educator and workforce trainer tasks from 15 percent to 54 percent.
   
7. Use AI to assess the workforce. AI tools can analyze workforce data to identify trends, measure productivity, and evaluate skill gaps. These tools offer organizations a more data-driven, effective, and insightful approach to understanding employee performance and skills. AI-driven assessments can also help organizations make more informed decisions about promotions, training needs, and talent development. By offering a more objective and comprehensive view of employee capabilities, AI-based workforce assessments can also help tailor development programs to individual needs resulting in a more engaged and empowered workforce. While workforce development professionals will still need to audit and validate the assessment tools prior to implementation, AI can save them time with the initial development of these tools.
   
8. Develop and adopt clear organizational AI governance policies. Employees using AI tools without proper oversight or understanding can expose sensitive data, create possibilities for poisoning and other attacks, and fail to account for the inexactness of results from many AI models. This can make an organization vulnerable not just to breaches or accidental sharing of proprietary information, but also to unwanted impacts on process outcomes due to both accidental and deliberate errors from the AI tools. Additionally, a misalignment of training data and operational use cases can lead to high error rates, which can affect the outcomes of workflows. These are all aspects of governance and are generally well documented. Despite this, many organizations lack clear AI governance policies, leaving employees to use tools without organizational approval and exposing the organization to further security risks. The absence of clear governance policies coupled with a workforce untrained in AI can present considerable vulnerabilities to an organization.
   
9. Craft a well-defined digital transformation strategy for workforce development. Such strategies enable organizations to align their workforce with evolving mission objectives and technological advancements. This will foster a culture of continuous learning and adaptation. Investing in professional development for the workforce as part of a digital transformation strategy can not only significantly enhance productivity and innovation but also prepare workers to adapt to the fast pace of technical innovations and mission aspirations. For instance, research from Bain highlights that organizations that prioritize workforce upskilling alongside digital initiatives see improved performance and competitive advantage. In the DoD context, digital transformation strategies are key to ensuring mission readiness.
   
10. Invest in comprehensive AI training for all 11 DCWF roles up front to save time in the long term by minimizing the need for ongoing corrections or retraining. Ensuring that the AI training is effective is crucial. Pre- and post-tests should be used to evaluate the knowledge learned during the training. Scenario-based assessments should also be put into effect to ensure that the knowledge can be successfully applied to their role.

These recommendations are based on our team’s advisory role with DoD mission partners over the past two years. The SEI's AI Workforce Development team combines both a workforce development academic research background and DoD and federal agency AI expertise to help organizations tackle any workforce talent challenges that they may be encountering. We encourage you to collaborate with our team so that we can learn more about those challenges and propose solutions.