Artificial intelligence (AI) is already helping employees at the Department of Transportation (DOT) work more efficiently, but building trust in the technology will be critical to scaling its use in the department and across the federal government, a senior agency official said on Thursday.

Speaking at the ServiceNow Government Forum, Anil “Neil” Chaudhry, senior advisor for AI at DOT, said the department is deploying AI in targeted ways that support mission priorities while demonstrating tangible value for employees.

“What I do on a day-to-day basis is work with my partners, my stakeholders, my industry counterparts, to talk about this concept of safety and affordability through innovation, action, and accountability, which means we need to be delivering now, not two years from now,” Chaudhry said.

Within that framework, DOT is applying AI in several mission-focused areas, including advancing transportation autonomy and improving supply chain efficiency.

“One is really about advancing autonomy in the United States. So, autonomous vehicles, anything autonomy, how do we integrate it seamlessly in our national infrastructure? It takes a lot of AI to do that,” Chaudhry said.

“Then, the other thing is about enhancing the use of artificial intelligence to really optimize our supply chains,” he said. “Because every dollar we save in our supply chains is $1 delivered back to the American public.”

Miguel Donayre, forward deployed solution architect for applied AI at ServiceNow, echoed that message, urging agencies to start by clearly defining the problem they want to solve.

“My favorite saying is, ‘just because you have a hammer doesn’t mean everything’s a nail,’” Donayre said. “It’s true with AI … you really want to look back and say, what are we trying to solve? That should be your baseline for anything AI.”

Building trust in AI

Finding the right use cases is a start, but Chaudhry said the real hurdle for expanding AI use, including generative AI, is in securing employee trust.

“The only way to scale there is trust,” he said.

To help build that confidence, Chaudhry described taking a “double-blind approach” that compares human-generated work with AI-generated work without reviewers knowing which is which.

“Once you get to the point where they can’t distinguish whether it’s coming from a large language model or a person, you can actually scale that piece up, and you don’t need to tweak it,” he said.

On the other hand, if teams don’t trust outputs generated with AI, Chaudhry said they will simply redo the work themselves.

“If the finance staff is saying, I don’t trust whatever this AI output that the procurement staff put together, they’re not going to use it,” Chaudhry said. “They’re going to redo the work, and all of a sudden, your governance has failed.”

Incremental gains for employees

Chaudhry said the most effective path to broader AI adoption is focusing on incremental improvements that make everyday work easier for federal employees.

“Really, what AI is about is, how can it help me do my job a little better today?” he said, adding, “Some of my peers now, including myself, what we do is we give our AI agents homework, and then we go home, and next morning, when we come in, we have some ready-made stuff.”

That incremental improvement approach can add up across large organizations, he said.

“AI, to me, is also about the power of fractionals,” Chaudhry said. “If everybody gets 1% better, if you have 100 people in the organization, that’s real change. That’s real work.”

Chaudhry also encouraged federal employees to experiment with AI tools on lower-risk tasks to discover where they can add value.

“It’s okay to experiment; you should experiment. And what you’ll find is the best way AI works for you,” he said. “Until you try it … you’ll never know. And if you’re able to do that, you will find your happy place in AI that makes you successful.”

When used thoughtfully, Donayre added, AI can become a powerful assistant for employees across government.

“I think it’s the best sidekick for anybody in the world today,” Donayre said. “It makes your job so much easier.”

Read More About
Recent
More Topics
About
Grace Dille
Grace Dille is MeriTalk's Assistant Managing Editor covering the intersection of government and technology.
Tags