This research explores integrating Large Language Models into financial decision-making by using ChatGPT to analyze and enhance loan officer assessments for credit default prediction.

• Compares AI-refined text with human-written assessments
• Reveals significant differences in text length and semantic structure
• Demonstrates how AI can augment traditional credit risk evaluation
• Explores implications for financial security and lending workflows

For security professionals, this research offers insights into how AI can strengthen risk assessment protocols while maintaining human oversight in critical financial decisions.

Unleashing the power of text for credit default prediction: Comparing human-written and generative AI-refined texts