LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation-assisted interpretation

Life Cycle Assessment (LCA) is a standard tool for quantifying environmental impacts, but its interpretation phase often fails to provide actionable, real-world strategic pathways. Researchers sought to bridge this gap by developing an AI-assisted framework that translates technical LCA findings into implementation-oriented roadmaps, accounting for technological, social, and policy uncertainties.

The authors developed a perspective-conditioned Retrieval-Augmented Generation (RAG) architecture. The workflow consists of three stages:

1. Scenario Anchoring: Defining system boundaries and decarbonization targets based on LCA results.
2. Perspective-Specific Retrieval: Using four distinct datasets—academic literature, industry profiles, public discourse (e.g., Reddit, Bluesky), and EU funding/policy documents—to query information through a RAG pipeline.
3. Neutral Synthesis: Integrating the retrieved, perspective-specific insights into a consolidated, auditable roadmap without allowing the model to introduce ungrounded external information.

The framework was tested using a hydrogen-enabled diesel reduction use case for an Italian apple production facility, utilizing GPT-5 nano as the reasoning engine.

The framework successfully generated an auditable 2030 roadmap for the case study. By separating data into distinct perspectives, the model provided nuanced insights: academic sources highlighted system-level trade-offs, industry data focused on deployment pathways, public discourse identified legitimacy and safety concerns, and EU policy data pointed toward funding accelerators. The structured, constrained synthesis effectively mitigated hallucinations while maintaining cross-domain diversity, demonstrating that AI can act as a disciplined assistant for strategic decision-making beyond traditional LCA reporting.