NGA taking cautious approach to AI adoption in human resources
As the National Geospatial-Intelligence Agency adopts artificial intelligence into HR workflows, the organization is taking a prudent approach to ensure its workforce doesn’t become overdependent on the technology.
“My biggest fear is that in five years, we’re going to lose a lot of expertise because we have automated so many of the things that have helped those individuals really understand their tradecraft, understand HR and the nuances and complexities and be able to grow,” Sasha Muth, deputy director of human development at NGA, said Tuesday during a panel the Workday Federal Forum, presented by FedScoop.
Muth’s concerns come as the intel agency explores how AI capabilities can be leveraged to modernize the organization’s HR processes — such as by speeding up mundane administrative tasks or using the technology to make more informed workforce decisions. The efforts are aligned with those happening at the Pentagon, which in recent months has begun pushing AI integration across the department.
NGA has charted a five-year transformation plan to improve its human capital management enterprise with AI that includes strategies to develop the organization’s workforce along with the technology, Muth explained. Part of that work entails nailing down the AI governance needed for HR modernization — which she described as a complicated situation.
“How do we enable the workforce to use AI decisions such as performance management and promotions? And how do we enable that and have those guardrails ensuring that we are not creating bias and that we are creating validation in the process?” she said.
Jay Harless, NGA’s director of human development, noted during the panel that while the organization is moving fast on adopting AI, it is not trying to oversaturate workflows with the technology.
He emphasized that while some key tasks can be done effectively with AI models, there are some missions within NGA that the technology will never be used for — such as final assessments and decision making.
“We are moving fast safely by distinguishing what should be automated, what should be augmented and what should be kept purely human,” Harless said. “There are some things that will always be purely human.”
