Text Speaks Louder: New Insights into Personality Assessment Using Natural Language Processing

In the ever-evolving field of personality psychology, recent advancements in technology have opened new doors to understanding human behavior. Our latest preprint, “Text Speaks Louder: Insights into Personality from Natural Language Processing,” dives into the interdisciplinary use of AI to predict and analyze personality traits based on textual data.

Our lab members, David Gallardo-Pujol and David Saeteros, in collaboration with Daniel Ortiz-Martínez, have bridged psychology and machine learning by applying explainable AI techniques to Natural Language Processing (NLP) models. These models assess personality traits using two major frameworks: the Big Five and the Myers-Briggs Type Indicator (MBTI). The study compares these typological and dimensional personality models through the analysis of datasets like Essays and MBTI.

We utilize advanced AI models such as BERT and RoBERTa, which process text data to detect personality signals with moderate to high accuracy. The results show that NLP models, particularly in the context of the Big Five, capture language patterns that align well with established personality theories. However, our study also highlights the limitations of the MBTI dataset, where biases—rooted in the participants’ self-knowledge of their own MBTI types—affect the models’ predictions.

Our findings indicate that while NLP can be a powerful tool for understanding personality, it also presents challenges, especially in balancing model performance with theoretical accuracy. The future of personality psychology will likely rely on further interdisciplinary research to develop more transparent, valid, and reliable personality assessments.

To learn more about this groundbreaking work, you can read the full preprint here.