A new study reveals that artificial intelligence (AI) systems are surpassing humans in emotional intelligence (EI) assessments.
Researchers from the University of Geneva (UNIGE) and the University of Bern (UniBE) tested six AI models, including ChatGPT-4, using standard EI tests. The AI models achieved an average accuracy of 82%, significantly higher than the human average of 56%.
The AI models also generated new EI test scenarios, which were evaluated by over 400 human participants. These AI-created tests proved to be as reliable and realistic as the original tests. Experts suggest that AI could be used in education, coaching, and conflict management, but caution against over-reliance.