Agencies’ rollout of artificial intelligence tools have garnered lots of attention, but officials working behind the scenes at the Defense Department say the vast majority of their day-to-day responsibilities focus on the unglamorous work of getting their people and processes ready for emerging tech tools.
“I think we often get characterized, in this cool kids mantra, that we show up and we get to do all the fun work,” Timothy Van Name, the deputy director of the Defense Digital Service, said Tuesday at an AFCEA DC conference. “[But] about 90% of our time on any given effort is spent doing all of the really hard crawling through the muck, working through data that’s in a ridiculous arcane structure, working through a code base that hasn’t been updated since the 80s.”
Peter Ranks, the Defense Department’s deputy chief information officer for information enterprise, said too much of the focus has fallen on adopting new tools, and not enough on using those tools to tackle fundamental challenges like making sense of the agency’s vast trove of data.
“If you’re thinking only about the technology and not about how it’s going to deliver outcomes in the government, you’re not bringing us the whole solution. And cyber products and data analytic products are my worst offenders in this space now,” Ranks said. “They’re just great sets of tools, but sooner or later, I have to get to an analyst who’s going to be able to integrate all this data together, and we have right now a surplus of data collection for cyber work but a shortage of analysts who can actually make sense of all that data.”
Leonel Garciga, the Army’s director of information management, said “tight-knit” groups within DoD provide DevSecOps and deliver solutions to warfighters quickly, but the challenge remains in scaling those pockets of excellence across DoD.
“Everybody’s talking about AI and machine learning. I’m like, ‘How about we just automate our workflows?’ How much time is that going to save before we start tackling some of these bigger challenges? And that’s where I’m sitting right now. As I look at scaling this within the Army, it starts boiling down to what things do we need to do from a blocking and tackling perspective to get that fundamental, foundational layer, to be able to have a real conversation about doing DevSecOps, or have a real conversation about AI and machine learning,” Garciga said.
But those workflow challenges also exist outside of the defense community. Giovanni Onwuchekwa, the director of the General Services Administration’s innovations division, said his agency helped overcome bureaucratic hurdles in procurement when it launched its IT Schedule 70 Startup Springboard back in 2016.
The program gives startups with less than two years of experience an alternative to Schedule 70. Onwuchekwa said that put hurdles in the way of experienced IT executives who left large companies to launch their own ventures.
“Some other really amazing tech executives and technologists leave their companies and want to start up a brand new startup. Those startups, at the time, could not get on Schedule 70, regardless of all the experience, knowledge, skills and the assets that they bring … And so we’re continually looking for ways to make it easy for the people that need to deliver the services to get access to the federal marketplace,” Onwuchekwa said.