The Digital Divide: Why Government Workers Are Struggling to Adopt Generative AI

As generative artificial intelligence sweeps across the private sector, transforming everything from software development to marketing, a distinct narrative has emerged: a productivity boom driven by machine-learning tools. In the corporate world, employees are increasingly leveraging large language models to automate mundane tasks, draft emails, and synthesize massive datasets. However, inside the halls of government—where the potential for such efficiency gains could translate into more responsive public services and higher-quality outcomes for citizens—the reality is far more fragmented. Many public-sector employees remain caught in a state of hesitation, unsure of how the technology applies to their specific duties, how to effectively wield these new tools, or, perhaps most crucially, whether they are permitted to use them at all.

This disconnect is the primary takeaway from recent research conducted by InnovateUS, a nonprofit learning initiative dedicated to upskilling the public sector. Having trained more than 250,000 government workers across all 50 states and the District of Columbia, the organization offers a unique vantage point on the digital transformation of the public workforce. The individuals they serve occupy a broad spectrum of critical roles, ranging from public safety and benefits administration to complex legal services and executive leadership. The organization’s mission is to provide state and local government employees with accessible opportunities to build digital competencies, ultimately aiming to help them deliver services that meet the evolving expectations of the public and restore a sense of trust in government institutions.

The challenges facing these workers are not necessarily born of a lack of ambition, but rather a lack of clarity and support. During the no-cost, nation-wide workshops and online training sessions facilitated by InnovateUS, public servants are provided a forum to trade experiences regarding what strategies yield results and where they encounter dead ends. These peer-to-peer exchanges serve as a vital laboratory for innovation. Participants have, for instance, explored the potential for artificial intelligence to streamline benefits delivery, an initiative discussed with New York City’s health commissioner. In other instances, they have learned from both current and former state-government leaders about the logistical and cultural hurdles involved in introducing AI to a workforce that is often historically risk-averse.

One illustrative example comes from a state leader in Colorado, who shared his experience using generative AI to overhaul the performance review process. By integrating AI into the workflow, he reported a significant improvement in the quality of feedback provided to staff, greater overall worker satisfaction, a measurable reduction in turnover, and, most notably, substantial time savings. For many participants in these workshops, such stories are transformative. They demonstrate that AI is not merely a theoretical construct discussed in tech headlines, but a practical tool that, when implemented thoughtfully, can solve chronic bureaucratic inefficiencies. These exchanges allow public professionals to learn from the real-world experiments of their peers and evaluate which strategies they might successfully adapt to their own local or state agencies.

Beyond the workshops, the digital learning environment provided by InnovateUS has enabled the organization to systematically listen to the experiences of the public-sector workforce. This ongoing dialogue serves as the foundation for the inaugural report from the organization’s Observatory of Public Sector AI. The findings, which are based on a massive dataset of 106,300 precourse survey responses and 67,002 postcourse responses collected between June 2024 and May 2026, paint a stark picture of the current state of AI adoption. The data reveals a substantial, and perhaps widening, gap between the optimistic headlines surrounding generative AI and the grounded, often cautious reality of everyday working life in government.

What keeps public workers from using AI

The data collected by the Observatory of Public Sector AI challenges the common stereotype of the government employee as an entity inherently resistant to change. Instead, the findings suggest that the barrier to entry is largely defined by a lack of institutional infrastructure and guidance. Among the respondents entering the foundation-level generative AI course, 51% reported that they had never attempted to use the technology in any capacity. A further 14% had experimented with the tools on a personal level but had not yet integrated them into their professional workflows.

While the private sector and major technology firms are engaged in a frenetic race to build and deploy autonomous agents, a significant portion of the public workforce is still grappling with fundamental questions. Many workers are searching for clarity on what AI actually has to do with their day-to-day job descriptions, why they are being encouraged to use it in the first place, and, perhaps most importantly, how to use it with the level of precision and security required in a government setting. Furthermore, the question of permission looms large; many employees remain uncertain about whether their agency’s policies allow for the use of these tools, or if doing so might inadvertently violate data privacy or cybersecurity protocols.

These findings complicate the narrative that government workers are simply "technophobic." When the data is broken down, the reasons for non-use are largely procedural and conceptual. Among respondents who were not utilizing generative AI in their roles, 28% stated that they did not see a clear application for the technology within their specific job functions. Another 23% cited a lack of technical knowledge regarding how to operate the tools, and 19% explicitly pointed to uncertainty regarding whether they were authorized to use AI for work-related tasks.

The confusion is not limited to frontline staff; it extends into the ranks of management as well. Managers are often the gatekeepers of innovation within a bureaucracy, yet they face the same—or sometimes greater—levels of uncertainty. Among the managers surveyed who were not using AI, 24% admitted they did not know how to implement it, and 22% remained unsure whether the technology was permitted under their current departmental guidelines. This creates a feedback loop of inaction: employees look to their supervisors for guidance and authorization, but those supervisors, lacking their own clear directives or training, are often unable to provide the necessary support.

As the public sector continues to navigate this transition, the data from the Observatory suggests that the path forward requires more than just access to software. It requires a systematic effort to educate leaders, clarify institutional policy, and demonstrate concrete, low-risk use cases that align with the specific mandates of government work. Without this structure, the promise of generative AI to improve public services remains locked behind a wall of caution, leaving many dedicated public servants on the sidelines of a technological shift that could otherwise redefine their capacity to serve the public trust. As these professionals continue to share their experiences and pilot new projects, the hope is that these small, localized successes will eventually coalesce into a broader, more confident, and more efficient public sector.