
When people feel embarrassed, they prefer talking to AI chatbots instead of humans.
But when they are angry, they still want to speak with a real person.
This finding comes from new research at the University of Kansas, which looked at how AI and human interactions affect people’s emotions, especially regarding sensitive topics like COVID-19 vaccines.
Many people have faced frustrating experiences when using AI chatbots for customer service.
However, this study shows that chatbots may be useful in situations where people feel embarrassed.
The researchers found that AI’s nonjudgmental nature makes it a comforting option when discussing personal or uncomfortable issues. On the other hand, when people feel angry, they prefer a human who can listen and respond in a more understanding way.
The research team, led by Professor Vaibhav Diwanji from the University of Kansas, wanted to explore how AI influences people’s decisions, especially during emotional situations.
They focused on the COVID-19 pandemic, which created both anger and embarrassment due to misinformation, vaccine debates, and social pressures.
The study involved 100 participants who were first asked about their attitudes toward COVID-19 vaccines. Then, they watched video clips designed to trigger different emotions—anger, embarrassment, or a neutral reaction. Some clips showed scenes of domestic violence or awkward social situations, while others showed calm nature footage.
To track their reactions, researchers used eye-tracking technology, which monitored participants’ eye movements and facial expressions to measure their emotions. Afterward, participants were randomly assigned either an AI chatbot or a human researcher to discuss vaccine-related information.
When asked about their experience, participants who felt embarrassed preferred talking to an AI chatbot, while those who felt angry preferred a human.
This suggests that AI chatbots can be useful in situations where people feel uncomfortable sharing personal information. However, when people are upset, they may want the emotional connection and understanding that only a human can provide.
Professor Diwanji explained that AI is becoming more sophisticated and widely used. While it won’t fully replace human interactions, it is important for businesses and health professionals to use AI wisely. The study suggests that AI chatbots should be designed to support people in sensitive situations, while human agents should be available for handling more emotional or frustrating interactions.
“In today’s world, AI is not just a trend—it’s a tool that should be used thoughtfully,” Diwanji said. “By understanding people’s emotions, we can create more personalized and human-like experiences with technology.”
This research, published in the International Journal of Human-Computer Interaction, highlights the growing role of AI in customer service, healthcare, and marketing.
As AI technology advances, businesses and researchers will need to find the right balance between automation and human interaction to better serve people’s needs.