Skip to main content

Trieste-Next-2026-IHRR-banner.jpg

[09/26] At Trieste Next 2026, IHRR Professor Fausto Guzzetti will discuss how artificial intelligence can strengthen natural-hazard decisions without replacing human responsibility.

On Saturday 26 September, IHRR Professor Fausto Guzzetti will take part in “Managing Natural Hazards between Artificial and Human Intelligence”, a public panel at Trieste Next 2026, the Festival of Scientific Research. Professor Guzzetti will be joined by Luca Bortolussi, Professor of Computer Science and Director of the AI Lab at the University of Trieste, and natural-risk project manager Andrea Spillatura. The discussion will be moderated by Stefano Parolai , Professor of Solid Earth Physics at the University of Trieste. Organised by the University of Trieste and the Data Science & Artificial Intelligence Foundation ETS, the panel forms part of the festival’s fifteenth edition, dedicated to the interactions between biological minds and artificial intelligence.

Professor Guzzetti will approach the discussion from both scientific research and his experience at Italy’s Civil Protection Department. Decisions during emergencies do not rest on a single expert or algorithm: they emerge from the collective assessment of evidence, legal responsibilities and the resources available for action. The hardest step is often turning limited, incomplete and probabilistic information into choices that are necessarily binary: evacuate or not, issue an alert or wait. Artificial intelligence can help by integrating heterogeneous monitoring data, detecting anomalies, comparing scenarios, updating maps and translating complex evidence into a clearer shared picture. Used well, it can reduce the cognitive burden on decision makers and leave more room for reflection and judgement. The discussion will also examine the limits of that promise. AI results must reveal the data on which they depend, how reliable they are, and the conditions in which they may fail. Otherwise, apparent precision can make a result harder to question without making it more trustworthy.

Professor Guzzetti’s central argument is that AI can improve forecasts and create a clearer shared understanding of an emergency. It cannot, however, negotiate competing interests, decide what level of risk a community should accept or take responsibility for the consequences. Not yet, at least.

Photo credit: AI-generated banner created with OpenAI’s image-generation tool, using an existing portrait of Fausto Guzzetti as visual references. The view of Trieste is AI-generated and does not reproduce a specific photograph.