The public sector urgently needs to empower its workforce to use AI safely. According to a recent report, 91 percent of public-sector and local government IT leaders acknowledge that the use of AI tools and agents outside of official oversight creates business and mission risk. At the same time, factors like budget restrictions, aging infrastructure and unclear governance frameworks make safe AI adoption easier said than done. Thankfully, we can lean on each other to figure it out. Peer cities can openly share what's working and what’s not when it comes to AI to learn from one another. This capacity for collaboration will be one of the public sector's biggest advantages as it navigates the hurdles of using AI safely. AI IS INEVITABLE; SAFE ADOPTION IS NOT It’s no secret that government IT budgets are tight. In many instances, the lion’s share of that money (up to 80 percent) goes toward maintaining legacy systems instead of adopting new technologies. County and municipal IT leaders know AI governance is crucial, but they have limited resources at their disposal. There’s also the challenge of shadow AI. The aforementioned report found that most public-sector IT leaders (96 percent) have encountered AI applications or agents being implemented by employees in non-IT functions. When employees use non-sanctioned AI tools or agents, sensitive government or citizen data can inadvertently leave an organization’s controlled environment. Furthermore, nearly every app employees use today, even everyday software, features some level of AI functionality. The problem is, these capabilities are often “turned on” or utilized without official approval, training or oversight. Despite these challenges, AI is here to stay as a core layer of the public sector’s technology stack. It’s increasingly playing a crucial role in keeping essential services up and running, from utility billing to permitting to public safety.