But AI could help less wealthy countries realize the dream of a smart city that is innovative, efficient and data-driven, urban experts said at the Fortune Brainstorm AI Singapore conference on Tuesday.

“AI levels the playing field,” said Cha-Ly Koh, founder and CEO of Malaysian data analytics company Urbanmetry.

Koh explained the four processes required to become a smart city: data collection, data analysis, decision-making and action. “Today, data collection can be done for a fraction of the cost using AI,” lowering the barrier to entry for cities in less developed countries. Data analysis has also become more accessible.

AI can also make data collection less intrusive. Algorithms can mask faces, windows, addresses and other identifying details from drone video footage to meet privacy requirements. “ISO standards can be built directly at the system level in terms of data encryption[and]sovereignty,” said Shaun Koo, CEO and co-founder of H3 Zoom.AI, which uses drones and AI to conduct construction inspections.

Ultimately, governments, businesses and stakeholders want “insights,” he said. AI algorithms can integrate some of this unstructured data to provide “actionable results.”

Data and planning

Joe Xia, CEO of Jidu, a self-driving car company owned by Baidu and Geely, cited his previous experience as a co-founder of Mobike, the Chinese bike-sharing service, as an example of how data can help with planning. Mobike used transportation data from buses, taxis and bicycles to determine the most efficient “last mile” public transportation solutions. That, in turn, helped cities in China remap their bus stops to improve public transportation efficiency.

But it’s not all roses and moonshine.

Koh said she wanted to curb “some of the enthusiasm” around AI and smart cities, particularly around data-driven decision-making. “Can that be done purely by AI in this part of the region? I think we’re a long way from that,” she said, as “cities are fundamentally political.”

A political issue is labor, as workers fear they will be replaced by automated technologies. Taxi drivers in the Chinese city of Wuhan, where Baidu is testing a fleet of 500 Apollo robotaxis, are calling for restrictions on their use. Apollo cars are “taking jobs from the grassroots,” one taxi company wrote in a letter reportedly sent to the Wuhan government in late June.

Still, Xia said it’s “a little too early” to worry about large-scale job losses from robotaxis. He also suggested that new technologies could lead to job creation in the long run: Automation allows companies to expand production and services more effectively, which in turn leads to more jobs overall. (Jidu and Apollo are targeting different products and markets, with the former focusing on assisted driving for individual consumers and the latter on fully automated robotaxis for institutional customers.)

Koh warned against seeing smart cities as something like SimCity, the famous video game series about city management.

“People tend to forget that people are also protesting,” she warned. “If we start monitoring everyone, there is a danger that we will take it too far.”

CEO Daily provides essential context for the news that business leaders across the globe need to know. Every weekday morning, more than 125,000 readers rely on CEO Daily for insights from across—and inside—the C-suite. Subscribe now.