The Department of State is finding that success with artificial intelligence (AI) depends less on advanced algorithms and more on disciplined data governance, a department technology leader said Thursday.
Inderpal “Rani” Virk, division chief for administration and logistics applications at the State Department, said that AI systems can only scale when agencies establish trust in their underlying data.
“AI scales where data is trusted, and that trust requires ownership and governance,” Virk said while speaking at the ServiceNow Government Forum in National Harbor, Md.
Virk noted that achieving trust in data requires agencies to confront several persistent governance challenges. Chief among them is defining clear responsibilities for agency data and having reliable data.
“AI forces agencies to decide who is actually accountable for their data,” Virk said.
She pointed to an AI bot that the State Department uses to automatically update records in the Federal Procurement Data System every week to reduce data errors.
“You can modernize systems quickly, but you can also only scale outcomes when the data underneath is reliable,” Virk explained. “… everyone runs … to move fast, but they forget that skipping readiness usually slows you down at the later-on stage,” Virk said.
“When you talk to people about governance, they think that it really slows them down,” she added. “In reality, it is actually what allows AI to scale safely.”
Virk said that in practice, the State Department is addressing governance challenges by focusing on data readiness before scaling AI initiatives. She emphasized that AI readiness isn’t a technical challenge; it’s a discipline.
“In other words, the real work isn’t building the models, it’s preparing the data,” Virk said.