Personality Prediction Using RNN Leveraging the Big Five Model and Its Application in Personality-Adaptive Chatbots

Abstract

This study proposes a method to predict users' personality traits, using Recurrent Neural Networks (RNNs) based on the widely used Big Five personality model, and its application to generative AI-based personality-adaptive chatbots. Text data from essays are used as dataset for the RNN model to predict these personality traits, showcasing the innovativeness of using RNN model to capture certain personality patterns from long, coherent texts instead of short posts on social media. The results show that the model can predict some personality traits effectively, but it has certain difficulties with others, which could be a potential area that may need further improvement. The study also apply the predicted personality traits into generative AI models to examine personalized chatbot responses. The findings demonstrate that after considering the predicted traits from the prediction model, generative AI models can generate more targeted responses according to the given traits. This research underscores the importance of adjusting chatbot responses based on users' personality traits, contributing to more empathetic and user-centered digital interactions.