A study involving 6,282 participants revealed that people express more empathy towards chatbots when they believe they are interacting with humans. Conducted by researchers at Jerusalem’s Hebrew University, the research, published in Nature Human Behaviour, demonstrated that emotional support from AI chatbots is generally rejected unless participants think they’re speaking to a human. Even though AI can effectively mimic empathetic responses, they were rated as less supportive compared to those attributed to humans. Participants also preferred the human-like interactions, as these were associated with more positive emotions. The findings suggest that while AI can analyze emotional states without fatigue, leading to seemingly empathetic responses, the perception of empathy largely hinges on whether users believe they are interacting with a human. This underscores the complexity of human-AI empathy and the challenges in recognizing AI’s capabilities and emotional nuances.