Emotional Intelligence 2.0: ChatGPT's Surprising Reactions

A recent study has shown that ChatGPT, one of the top AI language models, has anxiety-like behaviors when dealing with emotionally charged subjects like trauma or distressful events. Researchers have found that when ChatGPT is presented with such material, its outputs became more biased and unstable, sometimes resulting in racist or sexist responses. Interestingly, the research revealed that mindfulness-based relaxation methods can be used to counteract these effects, rendering the responses of ChatGPT more objective and unbiased. Although AI itself cannot feel emotions, its capacity to simulate human-like responses raises an eyebrow regarding its future use to understand human psychology and assist mental health professionals. Nevertheless, experts advise that AI should not be a substitute for professional mental health care. 

Source: Times of India