Federal Women Leaders Reflect on AI Guidelines, Inclusivity

Federal leaders discussed how they are applying ethical principles to artificial intelligence in their workflows and operations, they said at the Women Tech Leaders Summit in Washington, D.C. Wednesday.

“When we use AI, it’s not just AI. We make sure there’s human interaction at the beginning of the use, and then at the end, so there’s always a human that is responsible for the actions and the AI is kind of directing us in leads,” FBI Cyber Assistant Section Chief Stephanie Walker said at the event. “It’s not conducting investigations. … There’s a lot of safeguards that we put in place in order to make sure we’re using it responsibly.”

Federal agencies are undergoing widespread initiatives to harness AI capacities on the heels of the October White House executive order that directs agencies to develop AI ethically.

Generative AI is growing in popularity, and agencies are exploring its use for their operations. Department of Veterans Affairs AI Product Lead Dr. Kaeli Yuen said she sees it helpful for cutting down on tedious, manual tasks.

“Maybe a step up from the emails use case is ambient AI, medical scribes, which is something that VA has been prioritizing…thinking of the AI as a research assistant,” Yuen said. “The AI scribe is just like a human scribe, who might be a college student, or someone without a lot of experience who writes your note for you. As a physician, you are responsible for reading and editing and then signing off on that note. I think that framework of thinking of it as an assistant is a very helpful construct.”

Working in tandem, AI and human staffers can process data more quickly. At FBI, AI can process information that the bureau receives and brings it to a human agent.

“FBI receives thousands of tips every single day. And they come in through one funnel area. And all these tips are actively worked and looked at and evaluated for threat to life,” said Walker. “What if AI looks at that data and is able to flag it faster than a human can?”

AI is evolving quickly, making the need for stronger and clearer standards key for safer and more inclusive implementation.

“Almost every organization has some sort of AI guiding principles: trustworthiness, transparency — all of those are great. But when the rubber hits the road is where you’re going to take that and put it into action,” said Maximus Vice President of AI/Machine Learning Kathleen Featheringham. “Because you can say, ‘Yes, I agree. It should be transparent.’ But what does that really mean?”

AI also has implications for growing a diverse workforce. According to Department of Health and Human Services Chief Data Scientist Kristen Honey, AI can encourage more women to stay in technology careers.

“If you have a community of friends, network relationships, women stay in tech,” Honey said. “I think that networking and relationship building and strengthening those ties are ways to really increase diversity. And lastly — resources. You can’t expect great people to stay in a place if they can’t have that professional growth and scale some of the solutions or keep pioneering new ones.”

The officials also echoed recurrent views about how AI adoption is encouraging more non-technical people to work with and think about technology.

“Some of my best AI developers aren’t developers. They are writers because they’re inquisitive; they can ask the right questions,” said Energy Department CIO and Responsible AI Official Bridget Carper.

Walker encouraged other women to “never stop learning.”

“AI, emerging tech, quantum technology, it’s all coming rapidly,” said Walker. “I would say don’t necessarily think of yourself as an AI person or not an AI person. … Curiosity, continuous learning, all sorts of things like that will be essential to moving forward in this field.”