Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.
Agencies have gotten buy-in from their employees when it comes to automating rote tasks liked data entry. But it’s a tougher sell to those employees — and the public — to trust artificial intelligence tools for data-driven decision-making.
Some agencies, including the General Services Administration, IRS and Defense Logistics Agency expect to save tens of thousands of work hours through robotic process automation-powered bots for back-office functions.
There’s also room for automation grow, according to research from the Partnership for Public Service and IBM’s Center for the Business of Government. Up to a third of Treasury Department employees, for example, hold jobs with tasks that could become automated in the near future.
About 28% of employees at the Government Publishing Office and 22% of employees and the Securities and Exchange Commission could also see aspects of their work automated.
But that’s just the beginning. Federal Chief Information Officer Suzette Kent said the administration is looking “beyond the technical” challenges of artificial intelligence, and addressing some of the hurdles that arise when the technology reshapes the way the government does business.
“It’s easy to automate a manual task, but when you significantly change a business process, you have to look at human resources, you have to look at how we engage with customers. You have to look at how we collect data and what the regulatory environment looks like … we have to retrain a broader swath of employees,” Kent said Thursday at an IBM/Partnership event. “The work in the middle that really changes the business process and the service delivery is the tougher work.”
National Science Foundation CIO Dorothy Aronson, who also serves as the agency’s chief data officer, said her agency has tackled data and IT infrastructure challenges tied to several AI projects, but reskilling employees remains the biggest challenge.
“What do we want to do with the people who will be impacted? That was the absolute hardest part of the problem and we failed to solve it. Although we solved the rest, the main learning from that problem was we have to be more aggressive about fixing the people problem,” Aronson said.
Finding training opportunities
But in order to help reskill and retrain employees, NSF launched its Career Compass Challenge last November. The idea is to make an app that builds a resume or a personality profile for an employee, then to match them to government jobs that should exist in the future. The challenge also looks to flag potential training opportunities to help them reach that goal.
“Just having the vision and that conversation has been really liberating for us. Often we talk about the future as if it’s incredibly vague and we don’t know where we’re going,” Aronson said. “But the app gave us a substantial thing that people could look at and say, ‘If that’s the vision, how do I build to that vision?’”
Andrea Bright, the assistant secretary of Customs and Border Protection’s Office of Human Resources, said artificial intelligence could help analyze video feeds along the border, which would give CBP employees more time to carry out their law enforcement duties.
“We are moving towards automating some of that activity that allows the Border Patrol agents to do the work that they want to do,” Bright said.
But in order to reskill the workforce of the future, employees will need to know how to work with data. Last year, the Department of Health and Human Services launched an agency-wide AI and data science training program.
The program gives employees a crash course in data privacy and security. That’s especially relevant for an agency that ranks as one of the biggest sources of data among civilian agencies. More than 200 employees signed up for the first cohort of the program, which only had about 70 seats. HHS’ second class had more than 800 employees express interest.
HHS also pairs employees on the mission side of the agency with the IT side to help them identify potential AI use cases. Ed Simcox, HHS’ Chief Technology Officer, said that’s important, because not every problem needs an AI solution
“A lot of times I think AI is a hammer looking for a nail … People seem to gravitate toward shiny objects, even the folks that are mission-based and leaders inside of the mission space that hear the great things that we’re doing around AI across the federal government and they say, ‘I think we might need some of that,’” Simcox said.