Most agencies still do not have a clear picture of all the ways AI could transform their operations and services, but they have seen enough to realize that eventually it will be transformative. Here are some areas where agencies might focus their efforts at this stage of the journey.
Take the Grunt Work out of Research
Generative AI (GenAI) has already proven to be a time-saver at ONR, where many staff members spend lots of time reading and summarizing highly technical research papers, said Anthony Parrish, Chief Data Officer, Office of Naval Research. For example, he said, you can input 12 research papers on a given topic, and GenAI not only will summarize each, but also will identify common themes across those papers, much like a great research assistant. “It’s like the difference between trying to dig a hole with a teaspoon [vs.] using an excavator,” said Parrish.
Automate Processes From Start to Finish
Although GenAI can tackle a lot of different work quickly and effectively, it tends to be best at individual tasks. An emerging technology called agentic AI has the potential to deliver even greater efficiency by orchestrating tasks to automate a process from start to finish, said Karl Hermann, Manager, Solutions Engineering, State, Local and Education and Federal Government at Zoom. For example, many people use Zoom’s AI Companion to generate meeting summaries. But they also can set up the technology to create a meeting, automatically pull data that attendees need to review, and assign and track follow-up tasks or milestones after the meeting. “We definitely see people interested in that kind of process,” Hermann said.
Make Policies and Regulations More Adaptive
Agencies use the best available information to craft a given policy, but it still might not work out exactly as envisioned. The problem is that it often takes a long time to figure that out and make the necessary adjustments. But AI could change that, said Brian Campo, Deputy Assistant Commandant and Deputy CIO, U.S. Coast Guard. The idea is to identify the key metrics needed to measure the success of a policy, then use AI to collect and analyze that data on an ongoing basis and to recommend policy tweaks. “AI can enable us to do data-driven, continuously improving policy — if we put the right things in place,” Campo said.
Build AI Infrastructure That Scales
When building AI pilots, infrastructure is rarely the focus. But as agencies shift from pilot tests to enterprise deployments, they must think about scalability, said Keith Ober, Solutions Architect Manager at NVIDIA, and assemble an ecosystem of technology to deliver high-performing AI systems. NVIDIA describes that as an AI factory. Unlike a general-purpose data center, an AI factory is designed to meet AI’s specific computational needs. “That isn’t just a [graphics processing unit] chip. It’s not just a server in a data center. It’s not the network connection or storage platform,” Ober said. “It’s all of those working together to give you efficiencies at scale.”
Rehumanize Digital Engagement
People should be forgiven if they don’t find many digital services to be all that engaging. They find the app they need, provide the information requested and hopefully get the outcome they want — sort of like a digital vending machine, impersonal and expedient. AI, however, makes it possible to rehumanize government, said Luke Norris ,Vice President of Platform Enablement & Digital Transformation, Granicus. Over time, agencies have captured a lot of data about their constituents’ needs and preferences. Now, AI can help agencies use that data to inform how they interact with them. “We’re really seeing that when you harness that data…you can really start to make sense of who these people are to drive much more personalized experiences that ultimately build trust,” Norris said.
Get Everyone Up to Speed on AI
Given how quickly the AI market is evolving, people naturally have many questions about the technology: What are its capabilities? What are the security concerns? What’s the deal with data governance? At this stage, educating employees is even more important than identifying use cases, said Todd Schroeder, Vice President, Public Sector, at Databricks. He said it reminds him of where agencies were with cloud in its early days. The lack of understanding created fear and trepidation, making it difficult for employees to see the potential benefits of the technology. “If you don’t get this technology and the education to the front line, you don’t find the high-value use cases,” Schroeder said of AI.
This article appeared in our report, “How to Deliver on the Promise of AI.” To read more about how governments are putting AI into action, download it here:
Leave a Reply
You must be logged in to post a comment.