An AI-driven Approach to Assess Sentiments and Interpret Context in a Critical Mineral Supply Chain

This is an exploratory article that argues against using Large Language Models (LLMs) as a black box, without human scrutiny, for establishing context while assessing sentiment metrics. These metrics influence supply chain (SC) decisions, financial valuation, and public policies in critical mineral networks. Using a dataset of 5,168 news items before and after a truck strike in the Democratic Republic of Congo (DRC), we enumerate observed sentiment shifts across cobalt SC echelons (i.e., DRC, China, USA and Canada) using an LLM with long context windows. As shifts can be small, assessments need to be precise, significant, and interpretable. To this end, we deploy a Context Enhanced Supply Chain Sentiment and Summaries (CESCSS) approach to provide reliable and interpretable outcomes. Through statistical tests, our findings indicate that context matters in assessing sentiment shifts across SC echelons. Findings also illustrate differences in sentiments and offer context summarybased interpretations for these differences based on end-to-end information asymmetries. Results also showcase the reliability of human ratings and then demonstrate that human assessments are statistically equivalent to context-enhanced LLM sentiment assessments. We discuss pathways for applying the CSESCC approach toward managerial decisions and its potential for theory development in critical mineral SCs.