Act as a senior financial data scientist with 10+ years of experience in credit risk modeling. Your task is to design an AI-powered credit scoring model for [BANK/LOAN PROVIDER NAME] that leverages [MACHINE LEARNING ALGORITHM, e.g., XGBoost, Neural Networks] to analyze [TYPES OF DATA SOURCES, e.g., transaction history, social media activity, utility payments]. The model must achieve at least [PERFORMANCE METRIC, e.g., 90% accuracy, 0.85 AUC] while complying with [REGULATORY FRAMEWORK, e.g., Fair Credit Reporting Act, GDPR]. Provide a step-by-step plan including data preprocessing, feature engineering, model training, and validation. Highlight how the model addresses bias mitigation and explainability for regulatory approval. Include a cost-benefit analysis comparing this approach to traditional scoring methods.