Explainable AI for Psychological Profiling from Digital Footprints: A Case Study of Big Five Personality Predictions from Spending Data

AI can infer personality traits from financial behavior with measurable accuracy: The study analyzed 6,408 individuals' spending data and successfully predicted Big Five personality traits using machine learning, demonstrating that digital financial footprints contain psychological signals that algorithms can detect.

Explainability methods reveal both general patterns and individual-level decisions: The research shows that two types of explanations matter—global rules that identify which spending patterns predict personality across the population (e.g., certain purchase categories linked to specific traits) and local explanations showing why a particular person received a specific personality classification based on their unique transactions.

Transparency enables validation and improvement by domain experts: By making AI decision-making visible through rule extraction and counterfactual explanations, psychologists and other experts can verify whether the model's logic makes sense, identify potential biases, and suggest improvements—rather than treating the system as a "black box."

Confidence levels in AI predictions correlate with the amount of evidence used: The study found that when the model was more confident in its personality predictions, more financial features contributed to that decision, suggesting that explainability methods can help users assess prediction reliability.

Understanding how AI reaches conclusions about individuals is crucial for responsible deployment: The emphasis on both global and local explanations highlights that responsible use of psychological profiling AI requires transparency not just about what patterns the model learned, but specifically how and why it classified each individual person.