While artificial intelligence (AI) can provide plenty of opportunities for Federal agencies to advance their missions, adoption doesn’t come without challenges, officials said this week.
During ATARC’s Advancing AI in Federal IT Virtual Summit on May 14 , experts from the U.S. Navy, General Services Administration (GSA), Department of Labor (DoL), and the Office of Naval Research discussed the challenges associated with AI and machine learning (ML), and what agencies should focus on to adopt AI technologies.
“Making sure that we’re not only leveraging best practices, but really creating pieces that are reusable and reachable by a wide audience,” said Krista Kinnard, Branch Chief of Emerging Technology at DoL, describing what her agency is doing to implement AI solutions. “And I think that’s really how we’re thinking strategically about how this technology can really enable our organization, but also making it accessible to a wider audience, to not just solve these problems but really solve agency-wide problems.”
Kinnard said that DoL’s aim is to continue protecting workers’ rights and to not let AI biases impact that mission. Additionally, she said that agencies should focus on human-centered design for strengthening AI, and that Federal IT experts have a “responsibility to make our technology more accessible.”
Allen Hill, Acting Deputy Assistant Commissioner for Category Management at GSA, said that agencies should work to make the best use of AI, and stressed the importance of knowing where data is located and how to access it
“Cloud – and particularly a modern network infrastructure – is fundamental to adopting AI,” said Hill. “It’s very important for agencies to make sure they understand the datasets are accessible and know where they’re at.”
For the U.S. Navy, working on a literacy campaign to establish baseline knowledge for AI is important for getting people beyond the “magic” of the technology AI. Robert Keisler, Data Science and Analytics Engineering Lead at the U.S. Navy, spoke more about AI accessibility during the event.
“You have to have a workforce that understands all the different types of algorithms and the different types of approaches, but our biggest challenges really are getting folks to get beyond the magic of AI and understand how AI can be applied to their problems, or not, and really discussing that with our customers and our decision makers and our leaders in making AI real,” said Keisler.
