How artificial intelligence can increase trust in cities

December 12, 2024

Listen to This Article

Local leaders are always looking for new ways to build people’s trust in government—whether that means launching an ad campaign spotlighting the hard work of individual city employees, as they’ve done in Bratislava, Slovakia, or publishing a record of every ballot cast in an election, as they’re working to do in Boise, Idaho. At the same time, cities are increasingly embracing prudent experimentation with generative artificial intelligence.

Some may consider these conflicting efforts, especially given the privacy and ethical considerations that come with the new technology. But a growing number of cities are successfully experimenting with ways that they might be able to complement each other—and exploring how artificial intelligence more broadly could be a force for boosting government accountability and, in turn, bolstering residents’ trust in leaders. And what they’re finding is that, whether it’s about raising the bar for frontline service delivery, creating pathways for residents to better understand policy, or helping provide consistency in city hall messaging, AI’s use doesn’t have to clash with or test trust. Instead, it could help make it stronger.

Here’s how.

Raising the bar for frontline service delivery.

One of the best opportunities for AI-empowered trust-building might be day-to-day service delivery, according to Stephen Goldsmith, who is director of Data-Smart City Solutions at the Bloomberg Center for Cities at Harvard University and who, together with colleague Juncheng “Tony” Yang, recently published a paper on AI’s trust-building potential. The idea, he says, is that by improving the responsiveness of frontline service delivery, AI can help fuel a virtuous cycle of expectations met and residents being more likely to seek city help again in the future.

Some cities are especially focused on finding ways that artificial intelligence can free their teams from paperwork and other mundane tasks so that they can focus more on the people they serve. In London, for example, social workers are now experimenting with tapping the technology to transcribe sessions so that notetaking doesn’t interrupt their attention to clients. These kinds of frontline deployments, as Goldsmith—who previously served as mayor of Indianapolis and deputy mayor of New York City—and Yang explain, offer two potential benefits to trust and accountability. First: giving employees more time to zero in on the people they serve also means they’ll have more capacity to customize interventions and deliver better results. At the same time, they argue, AI applications can help managers improve feedback loops by producing “a clear, auditable trail of decisions and actions” to ensure new frontline interventions adhere to ethical guidelines.

Improving residents’ ability to understand—and monitor—official action.

Another way local leaders can use AI to build public trust is by making it easier for residents to engage in the policy process. And that starts with increasing the transparency of local government.

For instance, San Jose, Calif., has tested the use of AI to translate public meetings in real time, so that residents of every community can engage with the process. Likewise, Mayor Yan Zhao of the nearby city of Saratoga has used AI to produce easily accessible summaries of city council meetings, so more residents can be involved in shaping what their community looks like in the future.

But that’s just the beginning, Goldsmith says, describing the potential for cities to take further steps, such as training neighborhood organizations to use generative AI tools to better engage with local datasets—and, in turn, provide better oversight over local government.

“They don't need a coder. They don't need a sophisticated data fellow assigned to their board,” Goldsmith says. Instead, he argues, genAI tools that are already on the market can help everyone from individual residents to community leaders be more connected to local policymaking.

Making city communication clearer and more consistent.

Generative AI can also help cities build trust by serving as a tool for ensuring clear, consistent, and helpful messaging to residents.

City leaders in Las Condes, Chile, knew there was a growing frustration among residents who thought they were getting different answers to the same question—depending on what office they visited or person they asked. So the city—which has earned the Bloomberg Philanthropies What Works Cities certification for exceptional data capabilities—started by emphasizing a foundation of solid data, analyzing 100,000 calls to determine the most frequently asked questions. Then, leaders collected over 1,000 common questions, standardized responses to each, and built an internal generative AI tool, Sofia, to serve as a resource to help city employees be clearer and more consistent in how they communicate with residents.

Residents “trust now that the answer they are receiving is the right one,” Mayor Daniela Peñaloza tells Bloomberg Cities.

As Goldsmith points out, consistency isn’t the only reason this approach helps build trust. It also “allows the public-service worker to use her time to solve problems more quickly—and that has a trust-enabling aspect to it, as well.”

The impact of AI deployment on trust in government is still coming into focus. And the need for local leaders to ensure this technology doesn’t actively harm faith in institutions will remain. After all, AI doesn’t inherently deliver trust-building outcomes—even when deployments include a human layer of review or when staff make an effort to personalize responses developed using the technology. Still, where hesitation around the ethics of AI use has held municipalities back from experimentation, Goldsmith and Yang see serious potential not just to navigate those concerns, but also to improve confidence—in part by making cities more proactive.

“The most responsive cities and the ones that have the most trust from their residents are ones that identify and solve big problems—and little ones along the way—regularly, and in advance of them becoming crises,” Goldsmith says.

He adds, “The use of generative AI will allow them to preemptively solve problems, and therefore develop trust.”