Comparative Personality Assessment of Gemini and OpenAI Using MBTI and Big Five Tests

Agents4Science 2025 Conference Submission84 Authors

Keywords: AI personality, Gemini, OpenAI, MBTI, Big Five

TL;DR: This paper compares the personalities of the AI models Gemini and OpenAI using human-centric tests like MBTI and the Big Five to explore the challenges of attributing psychological traits to AI.

Abstract: This paper delves into the comparative personality assessment of two prominent AI models, Gemini and OpenAI, employing the Myers-Briggs Type Indicator (MBTI) and the Big Five personality traits assessment as frameworks. The primary objective is to scrutinize and contrast the responses of these AI models when subjected to these human-centric personality assessments, thereby illuminating the inherent challenges and potential pitfalls associated with attributing human-like characteristics and psychological constructs to artificial intelligence entities. The investigation encompasses a critical examination of the methodologies employed in adapting these established personality tests for AI assessment, addressing concerns regarding validity, reliability, and the interpretability of results. Furthermore, the thesis explores the philosophical and practical implications of such assessments, questioning the extent to which AI can genuinely possess traits analogous to human personality, and the potential for these assessments to inform AI development, human-AI interaction, and ethical considerations in the deployment of increasingly sophisticated AI systems. Ultimately, this work contributes to a broader understanding of the complex relationship between artificial intelligence and human psychology, offering insights into the limitations and possibilities of anthropomorphizing AI.

Submission Number: 84