Thousands of federal employees are likely to see their jobs and work requirements change as artificial intelligence is adopted at their agencies, according to a Partnership for Public Service and IBM Center for the Business of Government report released Feb. 28.
“AI is expected to revolutionize how government works. For one, AI could enable federal employees to focus on core responsibilities related to their agencies’ missions and spend fewer hours on administrative duties. They are likely to have more time to deliver services, interact with customers and perform other mission-related tasks. Should AI become pervasive in federal agencies, employees will need to enhance their digital and data literacy and learn how best to use the technology to work with citizens effectively,” the report said.
“AI is sure to change the composition of the federal workforce, creating new jobs related to managing AI systems or requiring critical thinking. Jobs based mainly on tasks that can be automated would likely be phased out, and employees would have to learn new or different skills for other jobs.”
The report estimates that approximately 130,000 federal employees will be directly impacted by the rise of AI, with some agencies seeing over a quarter of their workforce impacted.
Experts interviewed for the report disagreed on whether those changes would take place over decades or years, but the recent past has provided examples of how such work disruptions play out.
“The changes brought on by AI would not be the first substantial disruption of the federal workforce. In 1985, 19 percent of full-time federal employees held clerical positions. In 2017, they constituted just 4.3 percent in the workforce, according to Office of Personnel Management data,” the report said.
“During those years, desktop computers and other technologies automated many clerical tasks, and new employees were hired to deliver programs in newly created agencies such as the Department of Homeland Security. Whatever changes AI brings will not be immediate but an evolution that will play out over years and decades.”
Employees in their 50s are the most likely to see near-term impacts on their work from AI, according to the report, while employees in their 20s will see the lowest impact.
But “impacted” by AI does not necessarily mean that a job will disappear entirely, as such automated systems can often take care of dull and repetitious tasks, while freeing up employee time for more valuable pursuits.
“These are the so-called mundane tasks that almost no public employee wants to do,” said Kevin Desouza, a professor of business, technology and strategy at Queensland University of Technology in Australia, in the report.
Employees will, however, likely have to work on becoming more digitally literate to interface better with new AI systems.
“Everybody will be a bit of a data scientist in the future. It doesn’t matter if you are an HR person, an IT person or a business person,” said Dorothy Aronson, chief information officer at the National Science Foundation, in the report.
Federal leaders and personnel offices should therefore begin to prepare and train their workforces today for the changes AI will likely bring in the future, the report recommended.
The Office of Management and Budget should see AI in the context of a cross-agency priority goal and work with the General Services Administration to develop a team like U.S. Digital Service to focus on AI, according to the report, and the Office of Personnel Management should establish an AI occupational series.
“Leaders should communicate with employees early and often about the potential of AI to disrupt and alter their work. Leaders and managers should learn from early adopters of AI, such as the U.S. Coast Guard, NASA and the Department of Health and Human Services,” the report said.
“They should find out the extent to which the workday changed for employees, what types of agency work AI helped these organizations accomplish, which tasks were automated successfully, and what kind of work employees might start doing in place of current, repetitious tasks that AI could perform.”