Public-sector concerns over AI are lessening, survey says

Public-sector IT and program managers are less concerned about their employees having the right skills to use artificial intelligence than they were a year ago, according to survey data by Google Public Sector and GovExec’s Insights and Research Group, shared exclusively with Route Fifty.

Sixty-three percent of respondents said they are most concerned about having skilled staff, compared to 69% a year ago. Meanwhile, 58% said they are concerned about privacy, down from 71% in 2023, and 57% said they are worried about security, compared to 77% last year.

While those concerns about state and local governments’ AI use remain, they appear to be becoming less intense as managers learn more about the technology, its uses and how to make sure their employees can take full advantage of it. Chris Hein, Google Public Sector’s director of customer engineering, said that, combined with the drop in concerns about privacy and security, the decreasing worry about staff skills shows that the technology is slowly becoming demystified.

“As a lot of the generative AI craze got kicked off, there was a huge amount of uncertainty just because of the lack of exposure to what the tools were and how they were working, what they were working off, why it was hallucinating so much,” Hein said. “All those questions were tough for people to get their heads wrapped around, because it felt like it came out of nowhere for a lot of people that weren't deep in the AI industry already. Overall exposure to the tool sets have made it less of an overall point of consternation.”

States and localities have been focused on getting their employees trained, especially on generative AI, which promises to revolutionize how they work. Priorities for that training have included making sure terms are standardized and expectations are set about what the technology can and cannot do. The massive interest in training and workforce development has caught the attention of other vendors, too: ChatGPT creator OpenAI recently unveiled a government-focused version of its software designed for agency use.

AI use is only set to accelerate, too, based on Google’s survey findings. Of the 300 state and local employees in IT and program management roles who were surveyed, 94% said they foresee increased AI usage in their agencies over the next one to two years. Hein said that showed the level of excitement in the public sector about the tech’s potential. “There's so few things in this world where you can get nine out of 10 people to agree on something,” he said.

The survey data also suggested that agency leaders are starting to view AI as “being beyond just a chat bot,” Hein said. Almost half of respondents — 46% — said they are using AI to support predictive analytics for decision making, up from 37% a year ago. Forty one percent said they are using the technology to support AI data analysis, while 40% are using AI to automate repetitive processes.

The closeness of the numbers on use cases means that, so far, there is little “consensus around the great use cases,” Hein said, although that will likely always be at least somewhat the case as “different agencies have different needs,” he added. Hein also said he wants to see a “groundswell” in the next few years as more agencies wake up to the technology’s possibilities and deploy it effectively.

The conversation around the public sector’s use of AI comes amid a broader conversation about government efficiency, spurred by the federal Department of Government Efficiency and several state copycats. More than two-thirds in the Google survey — 67% — said they believe AI will improve their overall efficiency and productivity.

Florida Gov. Ron DeSantis was among the latest to create a state-level DOGE, with his executive order urging agencies to leverage cutting edge technology to identify further spending reductions and reforms in state agencies, university bureaucracies, and local governments that have increased their spending, to include identifying and returning unnecessary federal grant funding.”

Hein said AI can play a role in technological efficiency and can benefit both agency employees and residents who use government services. Hein cited the example of Covered California, the state’s health insurance marketplace, which uses AI to help verify residents’ documents and so speeds up approval times for consumers while making workers’ jobs easier.

“A lot of former modernization projects fell victim to the fact that it might have been running on newer technology, but what the constituent was getting was the same,” he said. “So, you didn't up level how it actually felt for the people paying the bill, and you just made it better in the back end, which is important, but it's hard to convey to people that it's important. Where I think AI really stands out is that it allows us to do both. It allows us to make it more efficient; you can get a ton more done.”

Hein said there is a long road ahead still for broad AI adoption, and the focus on privacy and security will remain, as well as determining solid use cases. But for those in state and local governments, there appears to be a groundswell of interest that is likely to be sustained.

“In the public sector, it's all about the mission,” he said. “What I think people have come to terms with, and are now getting excited by, is that ability to achieve the missions with technology as an enabler in a way that it never has been before.”