Core Projects - Environment, Climate, and Biodiversity

Sustainable consumption is a key strategy for climate mitigation, but traditional methods (such as generic information campaigns and one-size-fits-all nudges) are often ineffective at translating pro-environmental attitudes into behavior change. This project explores the potential of large language model (LLM)-based chatbots to promote sustainable consumption, focusing specifically on food choices. LLM chatbots, such as those based on GPT-like systems, have emerged as promising tools that can address both motivational and capability barriers by providing tailored advice on why and how to change consumption habits.

The chatbot developed in this project aims to guide users in making environmentally-friendly food choices by offering personalized support for reducing their dietary environmental impact. The system incorporates a psychologically informed design to maximize effectiveness, focusing on motivation (why to change) and capability (how to change). By simplifying decision-making processes, supporting planning, and diagnosing barriers, the chatbot seeks to bridge the intention–behavior gap at low cost.

The effectiveness of the chatbot will be assessed through participants' food selection behavior, measured by the share of low-impact items chosen, alongside secondary outcomes such as persistence in sustainable behavior at a one-month follow-up.

This research will provide valuable insights into how AI-powered chatbots can be utilized as scalable tools for promoting sustainable consumption, offering causal evidence on their impact and the factors that enhance their effectiveness. The results aim to inform both the use of conversational AI in policy interventions and the design of such systems to support climate-relevant behavior change.

Before the onset of the SARS-CoV-2 pandemic crisis, teleworking had been embraced as an occasional working pattern driven by evolving labor market trends. In recent times, the severe health emergency triggered by COVID-19 coupled with growing environmental and climate concerns, has propelled the widespread adoption of remote working solutions. This shift has reignited the academic and institutional debate on the implications of working-from-home (WfH) practices prompting a deep examination of the phenomenon. Previous research has delved into the consequences of WfH for workers discussing how flexible arrangements affect work-life balance, career prospects, or job satisfaction. Other studies unveil the contribution of teleworking to the resilience and performance of organizations. To grasp the net impact of such arrangements on the economy and society, the environmental footprint of remote working has emerged as a promising research topic. Despite increasing academic efforts, the link between teleworking and energy or emission savings remains an ambiguous issue. Extant evidence suggests that the reduction in carbon emissions from fewer commuting trips may be offset by more frequent and longer non-work trips and increased household energy usage. The variety of research methods and underlying assumptions exacerbates the overall complexity returning a nuanced picture. Furthermore, most studies conducted thus far rely on survey data collected before the advent of COVID-19, leaving room for assessing the implications and efficiency of teleworking configurations in a post-pandemic scenario. The project aims to explore whether and under what conditions WfH represents a viable strategy to reduce the emissions footprint of commuting workers, with a particular focus on the Italian context. Our approach leverages primary data on the commuting habits and energy consumption behaviors of both teleworkers and non-teleworkers. Exploiting a multidisciplinary approach, we provide a comprehensive assessment of the environmental impacts of teleworking.

Natural freshwater environments - such as rivers, lakes, and streams - support local economies and community wellbeing: they supply water for agriculture and industry, support tourism, and provide recreational opportunities like kayaking, swimming, fishing, and canyoning. Rivers are also biodiversity hotspots, hosting rich communities of plants and animals. This biodiversity delivers essential ecosystem services, including water purification, flood mitigation, natural control of harmful species such as mosquitoes, and the provision of resources like fish and pollinators. However, river ecosystems are fragile. Because they collect water from entire drainage basins, they can accumulate pollutants and are vulnerable to human pressures. To protect freshwater quality and sustain the economic and recreational benefits they provide, interventions such as channel excavation and the creation of ecological corridors are regularly implemented. Monitoring programmes are then needed to evaluate the effectiveness and sustainability of these actions.

Traditional biodiversity monitoring is costly, time-consuming, and reliant on expert knowledge—from field sampling and species identification to analysing ecological interactions and informing policy. As a result, comprehensive surveys are infrequent; for example, the last extensive fish monitoring of the Serchio River dates back to 2007.

Our project will test faster, more affordable methods for assessing natural biodiversity and ecological interactions, using environmental DNA (eDNA) — genetic material that organisms naturally shed in water - to detect species without direct observation or capture. By combining eDNA sequencing with advanced data analysis tools developed by IMT researchers, the project aims to infer ecological interaction networks. By improving the automation of biodiversity data collection and analysis, our project will contribute to increasing the coverage and reducing the costs of biodiversity monitoring, support timely management decisions, and strengthen the protection of freshwater ecosystems.