Search
Browse By Day
Browse By Time
Browse By Person
Browse By Policy Area
Browse By Session Type
Browse By Keyword
Program Calendar
Personal Schedule
Sign In
Search Tips
As the electric vehicle (EV) sector expands to meet global climate targets, the demand for critical minerals such as lithium, nickel, and cobalt has intensified. Corporate sustainability reports play an important role in communicating environmental and sourcing practices, yet concerns remain about greenwashing—where claims may be incomplete, vague, or selectively framed. Ensuring transparency in mineral supply chains is essential for both public oversight and informed decision-making. This study explores how artificial intelligence (AI), specifically, Large Language Models (LLMs) augmented by Retrieval-Augmented Generation (RAG) and structured prompt engineering, can assist in the analysis of sustainability disclosures. Using the 2023 sustainability report of a major U.S.based electric vehicle manufacturer as a case study, we assess how three prompting approaches, Input-Output (IO), Reflection-of-Thought (ROT), and Enhanced ROT, perform in identifying common indicators associated with greenwashing in the context of lithium sourcing. Results suggest that combining RAG with structured prompts improves the consistency and scope of text-based evaluations. Enhanced ROT prompts, which simulate multi-step reasoning and analysis by multiple perspectives, produce the most uniform outcomes across repeated trials. Commonly identified elements include the use of general language, limited third-party references, and absence of lifecycle data. This research contributes to the development of AI-supported tools for sustainability analysis. It demonstrates how advances in natural language processing can support the monitoring of corporate environmental reporting and assist policymakers, analysts, and researchers in identifying patterns relevant to supply chain governance and ESG reporting practices.