With excitement building around automated technologies and garnering support from policymakers, Federal agencies are looking to implement better data practices, cloud computing, and cultural shifts to increase the adoption of artificial intelligence (AI) technologies, Federal officials said this week.
Speaking at ATARC’s AI Working Group Launch event on July 15, panelists emphasized that better data practices need to be in place to fully take advantage of AI.
“One of the starting points for any AI project we work with is the data. If you don’t have the data, you’re really not going to be able to get too far,” said Alexander Measure, economist at the Bureau of Labor Statistics.
Agencies are taking steps to corral and analyze their existing data, while looking towards new processes to improve future collection.
“Too often, when we’re talking about finding data, we just run around for nine months searching for data and bring it all together when that nine months could have been better spent digitizing your workflows,” said Michael Kanaan, co-chair for artificial intelligence at the U.S. Air Force.
To get the computing power needed for AI, agencies are turning to cloud computing.
“What we found out is that it was really inexpensive for us to go to the cloud, launch certain services, and start to run parallel processing in cloud,” said Pamela Wise-Martinez, chief enterprise architect at the Department of Energy.
“One of our bots that we deployed for a compliance check process … only accesses publicly available information on public websites, with no access to PII … or IRS systems or data in this particular system. It lended itself to being a cloud solution – it was a great candidate for that,” noted Mitchell Winans, special assistant at the Internal Revenue Service.
The challenge of changing agency culture is one of the largest barriers to greater adoption of automated technologies.
“It’s no surprise to anyone in the room that culture is probably 90 percent of the [challenge],” said Nevin Taylor, senior advisor with the Presidential Innovation Fellows. “AI right now is a buzzword and a noun. I think we need to better understand what we’re talking about, or at least have some common reference point,” he added.
While agency cultures may need to evolve, speakers noted that they see support from policymakers and excitement around automated technology and digital skills.
“The DoD has a long history of treating foreign languages as something we value … but in 2019 computer languages like Python, Java, and C should be treated commensurately. This currently stands on the floor of the Senate through the [National Defense Authorization Act],” said Kanaan.