Unsupervised Feature Alignment: Ethical and Explainable Contrastive Approaches in Multimodal Artificial Intelligence Systems . eigmj 2025, 4 (9), 21-29.