What Government Can Learn from the Private Sector About AI

A new global research report reveals that public-sector employees are lagging behind their private-sector counterparts in both the use and understanding of generative AI.

The study, commissioned by SAS and conducted by Coleman Parkes Research Ltd., surveyed 1,600 organizations worldwide, including data managers, CIOs and IT directors in the U.S. public sector.

Government Technology analyzed the results to deliver three big takeaways that state and local governments should know about the state of generative AI now, and what it means for the future.


The SAS survey compared public-sector employees to those in various industries, including banking, energy, health care and insurance. A striking difference emerged: public employees demonstrated a lower personal understanding of generative AI.
Only 36 percent of U.S. public employees rated their understanding of generative AI and its potential business impacts as “well” or “completely,” compared to 52 percent of all U.S. respondents. Meanwhile, 22 percent of public-sector respondents reported using generative AI daily, compared to 28 percent overall. Finally, only 30 percent of public-sector respondents were very or extremely familiar with their organization’s adoption of generative AI, compared to 47 percent overall.

“It shouldn't be too surprising, that was not an aha moment at all,” said Jennifer Robinson, a global government strategic adviser at SAS.

She added that the disparity may not be a surprise, but a wake-up call to government agencies that the private sector is hitting the gas pedal on generative AI.

“In the public sector, we often look over to the private sector and say, 'oh, you know, we want to think a lot more like the private sector,'” she said. “We want to be innovative like the private sector. Well, they’re moving ahead.”


The survey revealed a significant difference in how public- and private-sector employees view synthetic data — information artificially generated to protect sensitive data.

While private-sector respondents expressed interest in using or were already using synthetic data to address challenges, their public-sector counterparts showed more hesitancy.

“Using synthetic data is going to enable us to be able to run great predictions and ‘what if’ scenarios, without having to compromise anybody's security,” said Robinson. “I wonder if maybe [public-sector workers] don’t understand synthetic data, maybe that’s a concern? Are they not aware of it, maybe they don’t understand the benefits of it, maybe they think it’s hard to do. Or maybe they don’t trust the accuracy of it.”


The SAS survey highlighted another significant gap: the lack of AI governance standards in the public sector. Only 52 percent of public-sector respondents reported having a generative AI policy, compared to 65 percent of overall respondents in the U.S.

Robinson emphasized the importance of such policies, not just for governing AI use but for fostering innovation and practical applications.

“If I’m in leadership for my municipality, I would suggest to my peers that they should be setting up policy for generative AI, and they should be explaining to their employees how they plan on using generative AI,” she said.

To explore the full results of the SAS study, visit its interactive dashboard.