AI, Get Me a Human!

The widespread support for Pope Leo XIV’s recent encyclical on artificial intelligence is striking, if not ironic. At a moment when trust in institutions, politicians, and the economy is at historic lows—particularly in the United States and Europe—one of the world’s oldest institutions has managed to capture public attention by warning that efficiency driven by technological advances cannot come at the expense of human relationships, judgment, and dignity. For countries, cities, and communities trying to navigate the AI transition, the Pope’s message provides an excellent reminder that AI governance is not just a technical challenge, but a moral, and, by extension, democratic one. 

In the United States, widespread backlash against AI workplace integration and data-center construction demonstrates the increasing public scrutiny of the impending economic, environmental, and societal implications of AI. Some 77% of Americans do not trust the government to use or regulate AI responsibly, so how governments manage this transition will have profound impacts on public trust and democratic resilience. While the AI adoption challenge spans all levels of government, the most promising place to start is at the local level, which retains relatively high levels of public trust. A 2025 Gallup Poll indicates that trust in local service delivery is closely correlated with confidence in national institutions. This suggests that cities should use AI to improve public services—but not at the expense of public confidence that officials understand and prioritize the people they serve. 

AI can clearly advance service delivery by making processes faster, simpler, and more accessible. Few people would choose hours at the DMV over renewing a license online, or manual tax filing over an AI-assisted process. But rushed or poorly designed adoption can quickly backfire. As governments extend AI beyond routine, efficiency-enhancing tasks, the question is not just what can be automated, but what should be. 

That question is becoming more urgent as governments begin using AI as a frontline point of contact with citizens. New Orleans, for example, is planning to use AI to answer calls to the city’s 311 city services line to “improve efficiency and better manage limited resources.” Yet for many, the lived experience of AI implementation feels anything but efficient. While automation can deliver some benefits, the effects are often more complex—introducing new demands on agencies, increasing errors in routine exchanges, and adding pressure on public employees. Tasks that were once straightforward, like resolving a service issue or getting information about public services, are now mediated through systems that, at least at present, struggle with nuance, urgency, and hallucinations, and can easily end with hitting zero repeatedly or pleading, “Get me a human.” 

This frustration reflects a disconnect between governments and the people they serve. Simply providing services is not enough; persistent public frustration with the DMV shows that user experience matters as much as outcomes. While institutions turn to automation to manage staffing and resource constraints, residents expect more than accurate answers— they look for clarity, discretion, accountability, and even empathy from public servants and officials. As Pope Leo suggested, “In an era that favors speed and fragmentation, the human person still yearns to receive care and recognition from attentive minds, kind words and hands capable of tenderness.” When AI substitutes for, rather than supports, human-to-human interactions, it risks weakening the relationships that fortify civic engagement and democracy, and exacerbating the long-standing problems with misinformation and social alienation that predate modern AI.

This dynamic is most consequential at the local level. Ongoing interactions with municipal offices, librarians, public schools, and social services shape residents’ views of democratic institutions. Trust can be built (or eroded) through encounters with public servants who exercise judgment, respond to individual circumstances, and demonstrate a willingness to listen. Substituting these exchanges with systems that cannot replicate these qualities risks reinforcing the perception that government is distant, unresponsive, or indifferent. 

None of this suggests that AI has no role in strengthening democratic governance. On the contrary, some of its most promising applications are those that remain largely invisible to the public but materially improve government performance. The city of Kortrijk, Belgium is using LLMs to expand multilingual access to city services to better serve their diverse population. The city of Tucson, Arizona, is using AI to analyze historical data on past water system failures, allowing officials to anticipate where pipes are most likely to break and to carry out preventative repairs and upgrades. In these cases, technology enhances capacity—without displacing human interactions that anchor trust—by reducing the routine or frustrating exchanges that can undermine it. 

The effectiveness of implementation depends on how and where these tools are applied. AI is most valuable behind the scenes, streamlining administrative processes, improving data analysis, and reducing backlogs. It can become far more problematic when it replaces frontline interactions that require judgment, empathy, and discretion[TD1] [AM2] [AG3] [TD4] . Efficiency gains achieved by automating frontline engagements may prove to be self-defeating if they undermine long-term trust and legitimacy.

The challenge for local governments is not whether to use AI, but how to use it in ways that strengthen democracy. Government is fundamentally relational; citizens should not have to beg an AI chatbot to transfer them to a human being. Decisions about where and how to implement AI should be guided by a clear understanding of which interactions can be safely automated without degrading service—and which require human presence to sustain trust. Preserving those human connections, as Pope Leo argued, is essential not only for effective governance, but also for the resilience of democratic systems. 

The views expressed herein are those solely of the author(s). GMF as an institution does not take positions.