A purple box placed on a folding table in Alamo Square in San Francisco could have been mistaken for a tissue box. However, dozens of secrets were hidden inside.
Oliva Tai was nearby, watching as passers-by stood perplexed by a sign that read: “Confessional: Everyone has a secret related to AI.”
Among the confessions about delegating too much work or using artificial intelligence to suggest a place for a first date, there were several in which it was admitted to having used it to write very personal messages: “I had the AI write my breakup message.” “I made ChatGPT go out with my boyfriend. All messages.” “I used AI to write my girlfriend’s grandmother’s 80th birthday letter.”
“I used AI to win an argument. Hahaha, not that I’m proud, but I won.”
Tai, an artist who has repeated the project more than a dozen times in the Bay Area, New York and Denver, said calling it a “confession” implied that people were admitting that using AI made them feel at least a little guilty. But they used it anyway.
As more and more people turn to a chatbot to compose emails or perform routine tasks, the question arises: is it a wrong step to use it for personal messages? Is it ever okay to use AI to write a greeting card or birthday message? And what about a love letter or wedding vows?
There are those who use AI to write personal messages, even if they feel guilty about it
As part of her PhD research in consumer behavior, Danielle Hass conducted hypothetical experiments to determine how people felt when using artificial intelligence to compose personal messages. He found that people who had already used artificial intelligence felt just as guilty as those who had not.
“Everyone felt guilty,” said Hass, a professor at Kennesaw State University in Georgia. “The driving force behind this relationship was a lack of honesty; people felt like they were manipulating others.”
“Even if people are honest about their use of AI, there are still those who view it in a negative light,” said Scott Claessens, a senior researcher at the University of Bristol in the United Kingdom.
Claessens says his research suggests that people who use artificial intelligence are perceived as lazier, less competent and less trustworthy.
He noted that people often already turn to complementary resources for social interactions, such as using pre-written birthday cards or poems in love notes. However, he added that there is something about the ease with which AI works that is unappealing to people.
“What we really want in our social relationships is for people to make an effort, because that is a sign that you are committed to the relationship,” he said.
How to know when it is not appropriate to use AI
Sarah Newman, who collaborated with Tai as director of art and education at Harvard University’s metaLAB, said the adoption of AI is happening at such a rapid pace that norms are likely to emerge about when it is appropriate to use it and when it is not.
He suggested that we be guided by our mixed feelings. If you imagine using AI to write your wedding vows, would you feel comfortable telling your partner that you have used it? If not, then don’t do it.
“Personal messages don’t have to be perfect,” Newman said. “Part of what makes these messages authentic and endearing is precisely that they are imperfect.”
Tai and Claessens stated that some situations could justify the use of AI on a personal level, such as in the case of people who are in a romantic relationship and use it to communicate better with their partner when they do not share the same native language. People with learning difficulties, such as dyslexia, may also turn to her for help.
Hass believes people are quickly accepting that AI can help them be more efficient in some work contexts, but he says he’s glad people feel guilty using it for personal messages.
“The purpose of guilt in this context is to show us our mistakes in interpersonal relationships,” Hass said. “It’s telling us, ‘Hey, this is where you’re supposed to behave like a human being.’”
This story was translated from English to Spanish with an artificial intelligence tool and was reviewed by an editor before publication.