ARIA: Artificial Intelligence for Sustainability Assessment

Evangelos Kallitsis , Gregory Offer, Jacqueline Edge

ARIA (Artificial Intelligence for Sustainability Assessment) is a Python package designed to streamline
the calculation of environmental impacts based on the life cycle assessment (LCA) framework.
It leverages Brightway2 (Mutel, 2017) as its core infrastructure, enabling robust LCA modelling
while automating several steps that traditionally require tedious manual effort. ARIA utilises
large language models (LLMs) to match foreground system inputs and outputs to the appropriate
background datasets based on user-defined instructions. Rather than painstakingly searching for
relevant processes in LCA databases such as Ecoinvent (Wernet et al., 2016), ARIA uses LLMs to
suggest possible matches and alternative search terms, making the flow mapping process faster and
less prone to errors. This automation frees LCA practitioners from manual inventory selection, one
of the most time consuming steps in LCA. Once activities are connected to inventory data, ARIA
executes impact assessment, drawing on multiple characterisation methods available in Brightway2
to calculate environmental indicators (e.g., global warming potential, acidification, water use). It
then provides convenient visualisation and plotting functionality, allowing users to quickly interpret
results. Overall, ARIA offers a modular, user-friendly Python interface for LCA, integrating
advanced capabilities in background inventory searching and automatic impact assessment to foster
reproducible and efficient sustainability assessment. The ARIA package is open-source and
available on GitHub.