Latest AI news, expert analysis, bold opinions, and key trends — delivered to your inbox.
In a surprising twist, researchers from the University of Geneva and the University of Bern have found that ChatGPT and other advanced AI models outperform humans on emotional intelligence tests — suggesting that AI may be better at reading and responding to emotions than we are.
Six AI systems — including GPT-4, Gemini 1.5 Flash, Claude 3.5 Haiku, Copilot 365, DeepSeek V3, and OpenAI’s o1 — were tested using standard emotional intelligence assessments.
These models had to choose the most emotionally appropriate responses to complex human scenarios.
On average, the AIs scored 81%, while human participants averaged just 56%.
GPT-4 didn’t just ace the test — it also generated new, valid emotional intelligence assessments on its own.
According to the researchers, this wasn’t just regurgitation from training data — the AIs appeared to reason through emotional concepts in real time.
No, AI doesn’t “feel” emotions. But this study shows that it can recognize and respond to emotional cues in ways that may rival — or even surpass — humans in certain settings.
That’s a big deal for industries like:
Mental health support, where empathy and responsiveness are critical
Customer service, where emotional tone can make or break interactions
Education, where emotionally aware feedback enhances learning
In short: AI’s emotional IQ is rising — and fast.