No one wants to talk to someone who sounds mean, whether they are a person or a chatbot. But research has found that going too far in the other direction is also highly undesirable. The study, conducted by a team of Northeastern researchers, found that personality compatibility plays an important role in how people perceive an AI chatbot.
The study found that people were generally happier to interact with a chatbot whose personality aligned with their own. People who consider themselves highly social responded better to chatbots that were more energetic in their tone and behavior, while the more reserved participants responded better to calmer and more straightforward bots. However, bots that were overly enthusiastic no matter the context were consistently received poorly by participants.
“Trust in AI is often discussed in terms of accuracy, reliability or transparency,” said Hasibur Rahman, a doctoral student who helped lead the study. “Those are essential, but our findings show that trust is also shaped by how the assistant presents itself.”
