People Respond Better To Chatbot Sans Human Qualities

As more and more jobs become automated, behavioral scientists are trying to understand if people prefer more or less interaction during product purchases online with the help of customer care bots.

A recent study conducted by Media Effects Research Laboratory affiliated to Pennsylvania State University's Institute for Cyber Science examined the interactions in various scenarios between chat-ots and potential customers, with and without a display image, to especially understand the terms of the personability expected. 

The research published in Computers in Human Behavior this April is titled “Humanising chatbots: The effects of visual, identity, and conversational cues on humanness perceptions.” A crowdsourced platform, Amazon Mechanical Turk, was used to interact with 141 participants who were paid to take part in a live chat on the pretext of purchasing a digital camera for a friend’s birthday.

The lead author of the study is Eun Go, a former doctoral student at Pennsylvania State University, who is now an assistant professor in broadcast journalism at Western Illinois University. As it turns out, people prefer chatting with bots who give standard responses and had more expectations out of talking with humanlike bots. 

"They’re low-cost and easy-to-use, which makes the technology attractive to companies for use in customer service, online tutoring and even cognitive therapy — but we also know that chatbots have limitations. For example, their conversation styles are often stilted and impersonal,” co-author S. Shyam Sundar said on the rising importance of chatbots and the need to study them.

Chatbot A woman chats with virtual software personality 'George' on a computer screen in London, 12 September 2006. George, who is 39 years old, single and light-hearted, is looking for friends on the internet. He has gifts -- the ability to speak in 40 languages and with 2,000 people at the same time -- and one quirk: he doesn't really exist. George is a piece of software, the best of the speaking "chatbots" or talking robots. ODD ANDERSEN/AFP/Getty Images

Three factors were tested in eight scenarios during the interactions with the participants who were informed about the changing communication patterns every single time: high versus low anthropomorphism, high versus low message interactivity, chatbot versus human interactivity. Responses were recorded after each scenario and each time a human being was interacting with participants to create the scenarios. The purpose was to assess the need or lack of personability while chatting with bots performing customer care duties.

The results were unexpected because when the participants were given a cue as to whether a human or a machine was at the other end, it affected their perception of the interaction to a large extent. People had more expectations and expressed disappointment after having chats with human avatars attached to bots due to a lack of nuance in communication and continuous back and forth dialogue.

It emerged from the findings that people preferred chatbots without human qualities better. This was in spite of monotone predictable responses, proving that conversing with Apple’s Siri may not be as satisfying as one might have thought. They were only surprised and amused with non-human bots, since they were not expecting anything in return from the interaction and were only hoping to get the order placed.

It was appreciated more if the chatbot without the human avatar just improved the style of communication to a small extent, rather than the conversation becoming more interactive with more prompt and detailed responses. Just acknowledging the message typed by the participant before providing a complete response was enough to impress the participants.

The researchers concluded that the only way to study this unique preference further was to monitor the same interactions in a laboratory setting since this study, which was handled online might not have all the answers.