GSA, VA Officials Discuss AI Advances in Agency Operations

AI is transforming the way Federal agencies complete mission objectives, and as the technology becomes more effective and efficient, agency leaders are seeking to utilize AI to reshape the future of government work.

One tool at the Federal government’s disposal to help map progress with AI is the General Services Administration’s (GSA) AI Transparency Score. The score is an assessment that agencies can use against objective criteria to determine how much risk their organizations are willing to take on in their AI project.

Hear insights from Deputy Federal CIO Maria Roat. Learn More

“We took incremental steps. We worked with industry, we worked with our Federal partners, we worked with people in academia to say: what is this actionable, tangible thing that people can use in the AI space, especially during this AI hype cycle,” GSA Director of AI Implementations Anil Chaudry explained during the ATARC IT Acquisition Virtual Summit on July 14. “And we said that each agency’s unique, each organization is unique, each IT or AI service provider is unique, so you can’t have a one size fits all.”

The AI Transparency Score, Chaudry said, isn’t a pass-fail grading system, and he reported that GSA has moved the project into working committees. He hopes in a couple years that there will be national and international standards aligned to AI transparency.

At the Department of Veterans Affairs (VA), the agency is taking a look at how to make AI more modular and scalable. VA’s AI Director Gil Alterovitz described how to cut out inefficiencies in scaling cutting-edge AI, and compared it to a restaurant making pick-up orders during the COVID-19 pandemic.

“When you think about cutting edge AI, many times it takes years for it to be developed and to lead to actual character, like that meandering experience within a restaurant,” Alterovitz said. “We’re looking at how do you cut that time? How do you allow for it to build to this kind of scalable and modular model? Essentially, [we are] truly exploring the idea of building immutable AI modules that are sort of ‘to go.’”

Jordan Smith
About Jordan Smith
Jordan Smith is a MeriTalk Staff Reporter covering the intersection of government and technology.

Categories

Recent