A significant focus was placed on decisions that can shape future heat risk. AI-driven grid optimisation can support more efficient energy systems, while AI-enhanced building energy modelling and retrofit optimisation can inform decisions around building performance. In transportation, AI-coordinated traffic signals, transit scheduling and route optimisation can improve the efficiency of urban mobility systems. These applications demonstrate how AI can contribute to broader urban planning and risk-reduction decisions rather than being limited to heat forecasting.
The presentation also explored AI-based heat mapping. Advanced models can generate more detailed, block-level representations of heat rather than relying solely on city-wide averages. This can help identify emerging heat islands and inform decisions around tree canopy, cool roofs and shading. Combining different sources of information can also support modelling of potential cooling interventions before they are implemented.
AI can further contribute during preparedness and response through forecasting, scenario modelling, energy-load management and decision support during periods of extreme heat. The value of these applications lies in connecting information to operational decisions and helping decision-makers understand where action may be most needed.
At the same time, the discussion highlighted important limitations. AI depends on reliable and accessible data, while fragmented information across institutions can restrict its effectiveness. AI infrastructure also has its own energy and cooling requirements, creating an additional consideration when assessing the overall sustainability of these technologies.
The presentation therefore positioned AI as an enabling capability across the urban heat risk cycle—supporting governance, investment, planning, preparedness and response—while emphasising that its effectiveness depends on strong data, institutional coordination and the ability to translate analysis into practical action.
