With the rise of conversational artificial intelligences, many people are using this new technology, not only for university or work, but simply to have a conversation.
However, a recent study conducted by Cornell University, points that people would rather talk to a human than to an artificial intelligence.
To demonstrate this, the researchers divided a group of people into 219 pairs, each pair was told to talk about a political issue via text message, given one of three conditions: Both pairs in a group could use a platform smart reply; only one person could use it; none of the participants could use the AI.
In the end, smart replies accounted for 14.3% of the messages sent.
The researchers found that the use of intelligent responses, however, increased communication efficiency, positive emotional language, and positive evaluations of communication.
Less connected to responses generated by an AI
But of course, according to the study, anyone who thought they were talking to someone whose responses were generated by a artificial intelligencewere perceived as less cooperative and felt “less connected”.
“I was surprised to find that people tend to evaluate it more negatively simply because they suspect that it is using AI to help compose text, regardless of whether it actually is. This illustrates the lingering general suspicion that people seem to have of AI.”, says the postdoctoral researcher Jess Hohensteinmain author of the article.
They then did a second experiment, where 299 pairs of randomly assigned participants were asked to discuss a political issue under one of four conditions: no intelligent responses; using Google Smart Answers; using smart responses with a tool AI different but with a positive emotional tone; use AI with a negative emotional tone.
After that, they discovered that those who used Google’s intelligent responses, or the tool that created optimistic texts, had a more positive tone. They believe this illustrates the benefits of AI in creating text that requires a more positive flow.
