Where to Use AI and Other Tips From CIOs on Emerging Tech

Artificial intelligence, machine learning and other associated technologies aren’t a panacea for the federal government, but strategic investments in those technologies are becoming almost mandatory for agencies that want to stay ahead of the technological curve.

“We’re using AI and machine learning in several areas,” said General Services Administration Chief Information Officer Dave Shive, speaking Monday at an event hosted by the Professional Services Council. “We’re increasingly using AI, pointing our AI algorithms out to our cloud boundaries ... and taking a look not only at the logs presented in those cloud environments but the actions of people as they interact with those cloud technologies.”

The combination of increased data storage capacity through cloud computing and sophisticated algorithms provides GSA a means to up its insider threat-tracking game, allowing the agency to look at how employees act on datasets over time and “flag when actions don’t make a lot of sense,” Shive said.  

Shive spoke alongside Health and Human Services CIO Jose Arrieta and Justice CIO Joseph Klimavicz, who shared how their respective agencies are employing emerging technologies to further their missions.

Arietta said HHS recently undertook three emerging tech pilots, including a recurrent neural network he hopes will improve the agency’s acquisition decision-making. For the pilot, Arietta said HHS technologists fed 9,000 contracting statements of work through AI algorithms using natural language processing. The algorithms scored each word in every sentence and “vectorized” those statements of work before they were fed into a predictive model based on reporting from the Government Accountability Office. The goal, Arietta said, was to “flag what may be inherently governmental and what may not be,” allowing for an “apples-to-apples comparison on price points.”

Arietta, who has been vocal about turning HHS into a testbed for emerging technology since he came into the job this summer, is also partnering with the Defense Information Systems Agency to test new ways to assess identity at the edge of networks.

“We’re leveraging emerging technology to push the envelope,” Arietta said.

Klimavicz said the Justice Department is testing 5G phones and beginning to look at quantum computing. While those technologies are on the near horizon—the Trump administration recently launched a committee to examine quantum efforts—Klimavicz said Justice is already making use of robotic process automation and machine learning to pair its growing datasets with analytic platforms.

Ideating within a bureaucracy isn’t easy business, Klimavicz said, so he spends a lot of time with industry technologies and with other government CIOs, likening it to a “CIO support group.” If they do something quickly at low costs, he said, it might work for him, too.

All three CIOs agreed that—in most cases—it makes far more sense for the government to buy commercial technologies rather than build them. In addition, they agreed that employees—existing or potential new recruits—must factor into tech decision-making processes for the future.

“If you do transformation and you don’t transform your workforce, you leave people behind—often your best people,” Shive said. “We retrain and reskill our staff and make sure they are educated. We go after new talent but we’re careful to maintain the legacy that exists. For those who think legacy IT people can’t become modern IT people, I say this: This new world we’re in, this is no different than you at the start of your careers.”