Mapping Research on Financial Inclusion and Credit Scoring: A Systematic Review with NLP-Assisted Thematic Analysis

This study develops a systematic literature review examining the relationship between financial inclusion, credit risk/credit scoring models, and algorithmic fairness in emerging economies. It is grounded in a theoretical framework that links access to credit with poverty and inequality reduction, as well as the role of artificial intelligence and machine learning in financial decision-making and their alignment with the Sustainable Development Goals. The search was conducted in Scopus following PRISMA guidelines and was complemented by Natural Language Processing techniques, bibliometric analysis, and science mapping applied to a final corpus of 63 articles. The results identify five thematic clusters that articulate approaches related to risk, financial inclusion, ethics and algorithmic fairness, microfinance and digitalization, evidencing a transition from traditional models toward data-intensive approaches. The study concludes that models based on artificial intelligence and alternative data offer opportunities to expand financial inclusion; however, they also introduce risks of bias and opacity. Therefore, governance frameworks and algorithmic fairness principles are required to balance predictive efficiency, equity, and sustainable development.