U.S. government investment in artificial intelligence has grown significantly in the last few years, as evidenced by the additional funding for AI research in President Biden’s 2023 fiscal budget.
With more than $2 billion allocated to the National Institute of Standards and Technology and the Department of Energy for AI research and development, it is clear there is a growing enthusiasm for the technology in government.
Driving the push to implement AI is the urgent need to address federal employee burnout. A recent study found that almost two-thirds of government employees are experiencing burnout, a much higher rate than seen in the private sector. Furthermore, almost half of respondents are considering leaving their government jobs within the next year due to increased burnout and stress.
One immediate solution to help this potential crisis is the responsible implementation of artificial intelligence. AI can mitigate the impact of burnout by removing repetitive and time-consuming tasks and streamlining processes, reducing the overall burden and repetitive nature of government work. However, effective AI investments demand more than just funding and technology.
Agencies must balance efforts to scale investment in AI while responding to the unique needs and challenges of the many diverse teams that make up the federal workforce.
Currently, there is a lack of cohesive guidance leading government efforts around AI. While organizations such as NIST have released basic AI framework for RMF, organizations without any AI experience may struggle to build the necessary foundation for a mature, agile AI posture. To lay the groundwork for a long-term AI strategy while generating short-term gains to support the federal workforce quickly, agencies must consider three guideline components of AI.
Ensuring acceptable levels of data maturity
Dedicated funding for AI is only one component of an effective AI strategy. Before implementing new technologies, agencies must start with their existing processes –– beginning with their level of data maturity.
If an agency does not have enough historical data to analyze, or the data they do have is not organized, implementing AI can create extra work on the front end for federal workers who could find themselves sorting through inaccurate or incomplete data processed by AI. For example, in response to this challenge, the Department of Defense stood up the Chief Digital and Artificial Intelligence Office (CDAO) to lead the deployment of AI across DoD, including the Department’s strategy and policy on data.
Once agencies realize a baseline of data maturity, they can pilot basic AI applications such as automating basic tasks, empowering agencies to gather high-quality data and provide analysis and insight around that data, providing the information needed to create a scalable AI roadmap that can integrate with other IT modernization technologies.
But having a roadmap alone is not enough to ensure that AI-driven technologies are useful for the federal workforce.
Understand federal employees’ needs
To implement AI that truly supports federal workers, agencies need to understand the main pain points and challenges facing federal employees. For most private enterprises looking to implement new technology, user experience surveys would be a core part of the pilot program to ensure an analytics-driven understanding of the technology’s successes and gaps.
However, although employee input is a crucial part of the AI planning process, government surveys are often expensive and can take months or years to consolidate into actionable data.
One way to combat this difficulty is utilizing existing AI to inform AI investments. For example, instead of sending out a survey where the results may take months to receive, an AI dashboard may provide a real-time view of overall work showing what areas need more support or automate a simple response survey where employees can provide input.
Using relatively basic AI to evaluate implementation allows agencies to gain insight into the needs of the workforce more sustainably and effectively than surveys, showing IT leaders where to implement AI for the most impact.
Using AI to enhance the employee experience
Once agencies have an AI baseline and understand worker needs, the last step to implementing employee-focused AI is creating a robust AI-empowered employee experience program.
There are many ways that AI can help agencies with experience management, from automating timesheets to streamlining business decisions. When AI is designed with these improvements in mind, AI’s tangible benefits support both broader organizational goals and the humans working to achieve them.
Scaling AI beyond pilot programs remains a challenge. One of the primary responsibilities of the CDAO is to develop processes for AI-enabled capabilities to be developed and fielded at scale across the defense space.
CDAO addressed this issue by selectively scaling only proven AI solutions for enterprise and joint use cases. Prioritizing proven solutions ensures that the AI they are implementing runs smoothly, is easy to use and –– most importantly –– is familiar to the workforce. As AI solutions become more sophisticated, agencies can continue to expand until they have a fully scalable AI network designed by and for humans.
Contrary to much of the conversation around AI, people are the most essential component of successful artificial intelligence programs. For implementation to be successful, agencies need a human-centered AI mindset. Following these three guidelines to create human-centered AI creates space for the federal workforce to be exponentially more creative, productive and ultimately more effective in furthering agency missions, equipping leaders to elevate the full potential of their teams.
Dr. Allen Badeau is the chief technology officer for Empower AI, as well as the director of the Empower AI Center for Rapid Engagement and Agile Technology Exchange (CREATE) Lab.