The term EQ, or emotional intelligence, was coined in 1990 by two American psychologists, John D. Mayer and Peter Salovey.
They defined emotional intelligence as "a set of skills that help understand and regulate one's own emotions and those of others to live effectively and achieve goals in life". EQ contributes to improving the quality of life in personal and professional relationships, family, education, employment, mental health, and many other areas.
Recognizing the importance of EQ, researchers assessed the emotional intelligence of AI through 5 tests similar to those used to evaluate humans. The results show that AI models scored significantly higher than human participants, with an average accuracy of 81% compared to the human average of 56%. AI demonstrated a strong cognitive empathy by understanding the causes of emotions, their consequences, and regulatory strategies.
Researchers indicate these results contribute to the growing evidence that AI models, such as ChatGPT, have capabilities at least equal to, or even surpassing, many humans in social and emotional tasks previously considered unique to humans.
Despite AI's advancements, many experts refute the claim that "AI surpasses humans". They argue that tests are often based on artificially controlled conditions and overlook human strengths in specific contexts. AI may score high on structured tests but struggle in real-life emotional situations, with nuances and rich contexts characteristic of human communication.
Therefore, over-reliance on AI, especially for emotional support, raises concerns about emotional dependence, privacy, and the decline of natural emotional skills.
In 2023, studies from psychology and AI journals concluded that AI has limited EQ capabilities, restricted to specific tasks. A review in Frontiers in Psychology indicated that AI can recognize and respond to emotional signals but lacks the depth, authenticity, and ethical foundation that humans possess.
Meanwhile, human EQ is more sophisticated and tied to life experiences. While AI can simulate emotional intelligence, especially in sentiment analysis, emotion recognition in voice, text, and facial expressions, it is believed to lack true self-awareness, genuine empathy, and contextual understanding of human experience. For example, AI can detect sadness in a voice but cannot feel or understand sadness like a human.
Thuy Linh (According to Psychology Today)
