AI Surpasses Humans in Emotional Intelligence Tests, Study Reveals

A recent study conducted by researchers from the University of Geneva (UNIGE) and the University of Bern (UniBE) has revealed that generative artificial intelligence (AI) systems, including ChatGPT, outperform humans in emotional intelligence (EI) tests. The team selected six large language models (LLMs)—ChatGPT-4, ChatGPT-o1, Gemini 1.5 Flash, Copilot 365, Claude 3.5 Haiku, and DeepSeek V3—and subjected them to five EI assessments commonly used in both academic and corporate settings. These tests presented emotionally charged scenarios to gauge the AI's ability to understand, regulate, and manage emotions effectively. One typical scenario involved a colleague stealing another colleague's idea and receiving undeserved praise. The correct response, according to human evaluators, was to talk to a superior about the situation. When the same tests were given to human participants, the LLMs scored significantly higher, achieving 82% correct answers compared to 56% for humans. This gap highlights the AI's capacity to navigate complex emotional landscapes and offer appropriate behavioral suggestions. The researchers also explored the AI's potential for creating new EI tests. Specifically, they tasked ChatGPT-4 with generating new emotionally charged scenarios. Over 400 human participants then completed these AI-generated tests, which were found to be just as reliable, clear, and realistic as the original tests that took years to develop. This capability suggests that LLMs not only understand emotional contexts but can also reason about them in a nuanced and sophisticated manner. Dr. Katja Schlegel, a lecturer and principal investigator at the Division of Personality Psychology, Differential Psychology, and Assessment at UniBE, explains, "The LLMs' performance underscores their ability to grasp emotional intelligence concepts and apply them consistently. This opens up exciting possibilities for integrating AI into fields traditionally dominated by human expertise." Marcello Mortillaro, a senior scientist at UNIGE's Swiss Center for Affective Sciences (CISA), adds, "Our findings show that AI can not only select the right answers but also create new, contextually relevant scenarios. This versatility could revolutionize how we approach education, coaching, and conflict resolution." The implications of this study are significant. If validated through further research, AI's enhanced EI capabilities could lead to its widespread adoption in various domains. For instance, in education, AI tools could assist teachers in developing students' emotional intelligence, providing personalized feedback and guidance. In coaching, AI could help professionals identify emotional triggers and develop strategies for better emotional regulation. Additionally, in conflict management, AI could facilitate more effective communication and resolution processes, especially in high-stress environments. However, the researchers emphasize the importance of expert oversight and integration. While AI can offer valuable insights and support, it should complement human judgment rather than replace it entirely. The ethical considerations of deploying AI in emotionally sensitive areas must also be carefully addressed to ensure that technology enhances human well-being without causing harm. The study, published in Communications Psychology, marks a crucial step in understanding the potential of LLMs beyond their traditional language tasks. It challenges the notion that emotional intelligence is uniquely human and opens doors to innovative applications of AI in fostering emotional intelligence and improving interpersonal skills. Industry experts and company profiles provide additional context and validation. Companies like Anthropic and Meta, which develop and deploy LLMs like Claude and Gemini, have been investing heavily in expanding the emotional and social competencies of their AI systems. This trend aligns with the growing recognition that emotional intelligence is a critical component of effective human-machine interaction in both personal and professional settings. The success of these AI models in EI tests reinforces the broader push towards creating more empathetic and context-aware AI, which could transform sectors ranging from healthcare to customer service. Overall, the study's findings highlight the rapid evolution of AI technologies and their increasing ability to handle nuanced, emotionally complex tasks. As AI continues to advance, the collaboration between human experts and AI systems will be essential in leveraging these capabilities for positive social impact.