What Does AI Recommend For Your City's Resilience Plan?

Can artificial intelligence (AI) help practitioners more comprehensively update resilience frameworks? Planners must design indicators that capture the multifaceted nature of urban resilience. These indicators should provide valuable insights into the interconnectedness of urban systems, enabling planners to make informed decisions and prioritize interventions effectively. Yet gaps in personnel and existing approaches can hinder the development and implementation of context‑specific indicators.

In "Artificial Intelligence for Extracting Key City Resilience Indicators: An Application to the Smart Mature Resilience Framework" (Journal of the American Planning Association, Vol. 91, No. 4), Blanca López‑Catalán, Victor A. Bañuls, Josune Hernantes Apezetxea, and Leire Labaka propose an innovative AI‑based methodology that synthesizes key topics on resilience and generates context‑dependent resilience indicators.

AI's Recommendations For Resilient Water Distribution

The authors tested their AI‑driven methodology through a collaboration with a seasoned resilience practitioner — a municipal water director in Seville, Spain. The indicator generation focused on a current priority of urban policy: efficient water use in consumption. The meta‑prompt integrated into the AI model included key phrases such as "collective engagement and awareness," along with context-specific information about Seville.

After evaluating the AI‑generated result, the water director modified the prompt. The generative AI was informed about the elements the practitioner viewed negatively, for example, the use of undesirable terms such as waste; the inclusion of graywater was suggested instead. Additional barriers in the initial indicator were also communicated, including the difficulty of measuring the population's level of trust.

Working with the water director, the researchers requested that certain technical concepts, such as pipe breakages, be integrated into the analysis. As a result, the AI generated a practical, data‑driven approach to assessing the water distribution network's ability to minimize losses and deliver water reliably during droughts or other stresses. The results and refinements can be viewed here (Google Gemini, 2024).

The city planner was also included to ensure that the indicators were not only theoretically sound but also practical, relevant, and implementable. The city planner reviewed and refined the generated indicators to ensure that they were relevant and accurately reflected the city's unique resilience needs.

Resilience Maturity Model

Keywords used in the AI prompting and analysis were identified through a systematic literature review and classified using content analysis. The resulting model, called the Resilience Maturity Model, is a matrix of maturity stages, dimensions, and subdimensions that categorize policies and city stakeholders. This matrix supports the generation of city‑specific urban resilience indicators.

AI provides a novel approach to generating resilience indicators, but uncertainties in urban resilience planning, such as data quality, model assumptions, and evolving risks, must be taken into account. Because AI often relies on historical data and predefined frameworks, its ability to anticipate emerging challenges is limited. Integrating expert insights can help address these uncertainties by incorporating the latest data and trends, including human dynamics in hybrid physical–virtual environments. Doing so would enhance the model's responsiveness to both current and future risks.

Localizing Indicators

The Resilience Maturity Model provides cities with a road map for operationalizing resilience. This process should account for each city's unique assets, which are integral to formulating urban policies and strategies. A selection of basic indicators, with the option to enhance them with more contextualized measures, enables planners to think globally and act locally.

The proposed methodology aligns resilience indicators with the city's vision and goals. Indicator selection and design play a crucial role in identifying vulnerable areas, monitoring progress, and encouraging stakeholder involvement in sustainability efforts. As a practical guide for analyzing and reconsidering themes and metrics to strengthen resilience, the method offers a dynamic, self‑reinforcing cycle of new indicators, feedback, and implementation.

Top image: Photo by iStock/Getty Images Plus/ Yuuji