Cities worldwide accelerated AI-driven infrastructure projects this week. In the United States, Austin’s city council approved plans to study the energy and water demands of AI data centers, which power tools like chatbots and image generators. A report on worker training and safety audits for these facilities is due by October 2025. Meanwhile, Boston-based Shawmut Construction reported success with its AI safety system that uses smartphone location data to alert workers about hazards like falling objects or unsafe zones.

In China, Yizhuang unveiled a cloud-based medical network linking hospitals across Beijing. The system uses AI for tasks like patient triage and appointment scheduling, with public hospitals serving as test sites for future expansions. Nigeria’s Lagos rolled out AI software for its bus network to optimize routes and reduce delays. This addresses severe traffic in a city where roads are often overcrowded.

The U.S. government took steps to support AI infrastructure by proposing 16 federal sites for new data centers. These facilities would provide computing power for advanced AI research, including climate modeling and materials science. However, experts at the AI and Cities Forum cautioned that agentic AI (self-directed systems) in construction could hurt small businesses. Algorithms might accidentally favor large firms with more data, creating unfair bidding processes.

Globally, cities are balancing AI innovation with fairness concerns. While tools like smart traffic systems and safety monitors show promise, leaders emphasized the need for rules to prevent AI from widening gaps between rich and poor communities.