Maybe artificial intelligence isn’t so intelligent after all.
A recent study led by Bingjie Liu, assistant professor in the School of Communication, suggests that using AI assistance to write messages to friends or significant others can be ineffective at developing personal connections. The study presented 208 participants with hypothetical scenarios of maintaining a relationship with a fictional friend, Taylor, and found that people tend to emphasize being present and thoughtful in a relationship, Liu said.
“My biggest takeaway would be, at least in the context of the relationship, efficiency may not be the most important thing,” Liu said. “It’s actually the amount of effort you put in the relationship.”
According to an article in the Journal of Social and Personal Relationships, the participants were given one of three different scenarios: giving advice, showing support or celebrating a birthday.
The participants wrote a personal message to Taylor about their given scenarios and received a generated reply. Each reply mentioned Taylor received assistance in writing the response from AI, another human or none at all.
Liu said the participants who received replies that admitted to using AI or human assistance were dissatisfied and found relying on other sources for relationship help to be off-putting.
“In this study, people find [AI] inappropriate to use for this purpose,” Liu said. “There are still social norms people have in their mind.”
While there is no clear way to determine if a person sent you a message using AI, Liu said people may do a Turing test — a series of questions a person asks to determine whether they believe an action was performed by a computer or a human — to get a better interpretation of the message.
Jesse Fox, associate professor in the School of Communication, said honesty plays a crucial role in maintaining a healthy relationship as it’s important for a person using AI in their relationship to be truthful about the originality of their messages.
“If people are using this, it’s very important that people are transparent about that and are divulging their sources and inspirations,” Fox said.
When it comes to using technology that still has plenty of unknowns, such as AI, Liu said it is best to be careful with how people use it in their daily lives.
“I think people hopefully can adopt a more sophisticated way to think about the use of AI, when it’s right and when it’s not ok to use, and engage more in the process of perspective taking,” Liu said. “This is about how the recipient of a message perceives that message and their relationship.”