Supporting Sustainable Diet Policies by Knowledge-based AI
Every year ASI awards seed money to promising interdisciplinary projects. "Supporting Sustainable Diet Policies by Knowledge-based AI" is one of four ASI seed money winners 2024.
CO2 emissions are now a key societal concern. Food has a strong impact on the planet and has been linked to increased CO2 emissions (Henriksson et al., 2021). While many theories of climate change and sustainability exist, there is still a lack of comprehensive studies and tools that can inform policymakers and enable them to make effective decisions.
Within the context of food policies, part of the challenge is linked to estimating the footprint of food’s life cycle – with stages including food production, composition, supply chains, recycling, and more. The data for each of these aspects largely exists, but it is fragmented - it is unclear how the puzzle pieces fit together, making decision-making by policymakers difficult. The data is also dynamic, with strong contextual dependencies on time, location, and cultural values. Finally, there is no single best option, as the best course of action is often relative to the specific goals of the decision maker and the specific circumstances of the situation.
This project aims to support policymakers in making decisions about sustainable diet policies through interdisciplinary research that combines state-of-the-art sustainability theories and resources with knowledge-based artificial intelligence (AI) technologies. As there is a lack of integration between research studies in life cycle assessment (LCA) and knowledge technology applications in computer science, this project bridges this gap, enabling LCA insights and theories to be scaled up into a comprehensive framework, enabling knowledge technologies to make a societal impact and advance in the light of sustainability requirements (Allen and Ilievski, 2024).
The application will be powered by a comprehensive diet footprint knowledge base that builds on our expertise to identify, extract, and consolidate relevant information. To make the resulting knowledge graph available to users, a user-friendly interface will be developed. For a given food product corresponding to a packaged item or a recipe, the interface will provide analytics about its carbon footprint based on information about its ingredients.
Research Team
- Dr Filip Ilievski, Assistant Professor, Faculty of Science, Artificial intelligence
- Dr Reinout Heijungs, Associate Professor, School of Business and Economics, Operations Analytics