Illustration: Raven Jiang\/The Guardian<\/p>\n
Earlier this spring, Nik Vassev heard a high school friend\u2019s mother had died. Vassev, a 32-year-old tech entrepreneur in Vancouver, Canada, opened up Claude AI, Anthropic\u2019s artificial intelligence<\/a> chatbot.<\/p>\n \u201cMy friend\u2019s mom passed away and I\u2019m trying to find the right way to be there for him and send him a message of support like a good friend,\u201d he typed.<\/p>\n Vassev mostly uses AI to answer work emails, but also for personal communications. \u201cI just wanted to just get a second opinion about how to approach that situation,\u201d he says. \u201cAs guys, sometimes we have trouble expressing our emotions.\u201d<\/p>\n Claude helped Vassev craft a note: \u201cHey man, I\u2019m so sorry for your loss. Sending you and your family lots of love and support during this difficult time. I\u2019m here for you if you need anything \u2026\u201d it read.<\/p>\n Thanks to the message, Vassev\u2019s friend opened up about their grief. But Vassev never revealed that AI was involved. People \u201cdevalue\u201d writing that is AI-assisted, he acknowledges. \u201cIt can rub people the wrong way.\u201d<\/p>\n Vassev learned this lesson because a friend once called him out for relying heavily on AI during an argument: \u201cNik, I want to hear your voice, not what ChatGPT<\/a> has to say.\u201d That experience left Vassev chastened. Since then, he\u2019s been trying to be more sparing and subtle, \u201cthinking for myself and having AI assist\u201d, he says.<\/p>\n Since late 2022, AI adoption has exploded in professional contexts, where it\u2019s used as a productivity-boosting<\/a> tool, and among students,<\/a> who increasingly<\/a> use chatbots to cheat<\/a>.<\/p>\n Yet AI is becoming the invisible infrastructure of personal communications, too \u2013 punching up text messages<\/a>, birthday cards and obituaries<\/a>, even though we associate such compositions with \u201cfrom the heart\u201d authenticity.<\/p>\n Disclosing the role of AI could defeat the purpose of these writings, which is to build trust and express care. Nonetheless, one person anonymously told me that he used ChatGPT while writing his father of the bride speech; another wished OpenAI had been around when he had written his vows because it would have \u201csaved [him] a lot of time\u201d. Online, a Redditor shared that they used<\/a> ChatGPT to write their mom\u2019s birthday card: \u201cShe not only cried, she keeps it on her side table and reads [it] over and over, every day since I gave it to her,\u201d they wrote. \u201cI can never tell her.\u201d<\/p>\n One person anonymously said that he used ChatGPT while writing his father of the bride speech. Illustration: Raven Jiang\/The Guardian<\/p>\n Research about transparency and AI use mostly focuses on professional<\/a> settings<\/a>, where 40%<\/a> of US workers use the tools. However, a recent study<\/a> from the University of Arizona concluded that \u201cAI disclosure can harm social perceptions\u201d of the disclosers at work, and similar findings apply to personal relationships.<\/p>\n In one 2023 study<\/a>, 208 adults received a \u201cthoughtful\u201d note from a friend; those who were told the note was written with AI felt less satisfied and \u201cmore uncertain about where they stand\u201d with the friend, according to<\/a> Bingjie Liu, the lead author of the study and an assistant professor of communication at Ohio State University.<\/p>\n On subreddits such as r\/AmIOverreacting<\/a> or r\/Relationship_advice, it\u2019s easy to find users expressing distress upon discovering, say, that their husband used ChatGPT to write their wedding vows<\/a>. (\u201cTo me, these words are some of the most important that we will ever say to each other. I feel so sad knowing that they weren\u2019t even his own.\u201d)<\/p>\n If I heard that you were sending me an email and making it sound more empathetic than you really were, I wouldn\u2019t let it go<\/p><\/blockquote>\n Dr Vanessa Urch Druskat<\/p>\n AI-assisted personal messages can convey that the sender didn\u2019t want to bother with sincerity, says Dr Vanessa Urch Druskat, a social and organizational psychologist and professor specializing in emotional intelligence. \u201cIf I heard that you were sending me an email and making it sound more empathetic than you really were, I wouldn\u2019t let it go,\u201d she says.<\/p>\n \u201cThere\u2019s a baseline expectation that our personal communications are authentic,\u201d says Druskat. \u201cWe\u2019re wired to pick up on inauthenticity, disrespect \u2013 it feels terrible,\u201d she says.<\/p>\n But not everyone draws the same line when it comes to how much AI involvement is tolerable or what constitutes deceit by omission. Curious, I conducted an informal social media poll among my friends: if I used AI to write their whole birthday card, how would they feel? About two-thirds said they would be \u201cupset\u201d; the rest said it would be fine. But if I had used AI only in a supplementary role \u2013 say, some editing to hit the right tone \u2013 the results were closer to 50-50.<\/p>\n Using AI in personal messages is a double gamble: first, that the recipient won\u2019t notice, and second, that they won\u2019t mind. Illustration: Raven Jiang\/The Guardian<\/p>\n Using AI in personal messages is a double gamble: first, that the recipient won\u2019t notice, and second, that they won\u2019t mind. Still, there are arguments for why taking the risk is worthwhile, and why a hint of AI in a<\/a> Hinge message<\/a> might not be so bad. For instance, AI can be helpful for bridging communication gaps<\/a> rooted in cultural, linguistic or other forms of diversity.<\/p>\n Plus, personal messages have never been totally spontaneous and original. People routinely seek advice from friends, therapists or strangers<\/a> about disagreements, delicate conversations or important notes. Greeting cards have long come with pre-written sentiments (although Mother\u2019s Day founder Anna Jarvis once scolded<\/a> that printed cards were \u201clazy\u201d).<\/p>\n Sara Jane Ho, an etiquette expert, says she has used ChatGPT \u201cin situations where I\u2019ve been like: \u2018Change this copy to make it more heartfelt.\u2019 And it\u2019s great copy.\u201d<\/p>\n Ho argues that using ChatGPT to craft a personal message actually shows \u201ca level of consideration\u201d.<\/p>\n Graphic with three lines of text that say, in bold, \u2018Well Actually\u2019, then \u2018Read more on living a good life in a complex world,\u2019 then a pinkish-lavender pill-shaped button with white letters that say \u2018More from this section\u2019<\/a><\/p>\n Expressing sensitivity helps build relationships, and it makes sense that people who struggle with words would appreciate assistance. Calculators are standard digital tools; why not chatbots? \u201cI always say that the spirit of etiquette is about putting others at ease,\u201d she says. \u201cIf the end result is something that is nice for the other person and that shows respect or consideration or care, then they don\u2019t need to see how the sausage is made.\u201d<\/p>\n I asked Ho what she would say to a person upset by an AI-assisted note. \u201cI\u2019d ask them: \u2018Why are you so easily offended?\u2019\u201d Ho says.<\/p>\n Plus, she says using AI is convenient and fast. \u201cWhy would you make yourself walk someplace if you have a car?\u201d she asks.<\/p>\n Increasingly, people are drifting through digitized lives that reject \u201cthe very notion that engagement should require effort\u201d, at perceiving less value in character building and experiences like \u201cworking hard\u201d and \u201clearning well\u201d, author and educator Kyla Scanlon argued in an essay<\/a> last month. This bias toward effortlessness characterizes the emotional work of relationships as burdensome, even though it helps create intimacy.<\/p>\n \u201cPeople have sort of conditioned themselves to want a completely seamless and frictionless experience in their everyday lives 100% of the time,\u201d says Josh Lora, a writer and sociologist who has written about AI and loneliness<\/a>. \u201cThere are people who Uber everywhere, who Seamless everything, who Amazon everything, and render their lives completely smooth.\u201d<\/p>\n Amid this convenience-maxxing<\/a>, AI figures as an efficient way out of relational labor, or small mistakes, tensions and inadequacies in communication, says Lora.<\/p>\n We use language to be understood or co-create a sense of self. \u201cSo much of our experience as people is rendered in the struggle to make meaning, to self actualize, to explain yourself to another person,\u201d Lora says.<\/p>\n But when we outsource that labor to a chatbot, we lose out on developing self-expression, nuanced social skills, and emotional intelligence. We also lose out on the feelings of interpersonal gratitude that arise from taking the time to write kindly to our loved ones, as one 2023 study<\/a> from the University of California, Riverside, found.<\/p>\n Many people already approach life as a series of objectives: get good grades, get a job, earn money, get married. In that mindset, a relationship can feel like something to manage effectively rather than a space of mutual recognition. What happens if it stops feeling worth the effort<\/a>?<\/p>\n Summer (who requested a pseudonym for privacy), a 30-year-old university tutor, said she became best friends with Natasha (also a pseudonym) while pursuing their respective doctoral degrees. They lived four hours apart, and much of their relationship unfolded in long text message exchanges, debating ideas or analyzing people they knew.<\/p>\n If the end result is something that is nice for the other person … then they don\u2019t need to see how the sausage is made<\/p><\/blockquote>\n Sara Jane Ho, etiquette expert<\/p>\n About a year ago, Natasha began to use ChatGPT to help with work tasks. Summer said she quickly seemed deeply enamoured with AI\u2019s speed and fluency. (Researchers have warned the technology can be addictive<\/a>, to the detriment of human social engagement.) Soon, subtle tone and content changes led Summer to suspect Natasha was using AI in their personal messages. (Natasha did not respond to a request for comment.)<\/p>\n After six years of lively intellectual curiosity, their communication dwindled. Occasionally, Natasha asked Summer for her opinion on something, then disappeared for days. Summer felt like she was the third party to a deep conversation happening between her best friend and a machine. \u201cI\u2019d engage with her as a friend, a whole human being, and she\u2019d engage with me as an obstacle to this meaning-making machine of hers,\u201d Summer tells me.<\/p>\n Summer finally called Natasha to discuss how AI use was affecting their friendship. She felt Natasha was exchanging the messy imperfections of rambling debate for an emotionally bankrupt facsimile of ultra-efficient communication. Natasha didn\u2019t deny using chatbots, and \u201cseemed to always have a reason\u201d for continuing despite Summer\u2019s moral and intellectual qualms.<\/p>\n \u2018AI is unable to give meaning to something because it\u2019s outside of the semantics produced by human beings,\u2019 says philosopher Dr Mathieu Corteel. Illustration: Raven Jiang\/The Guardian<\/p>\n Summer \u201cfelt betrayed\u201d that a close friend had used AI as \u201can auxiliary\u201d to talk to her. \u201cShe couldn\u2019t find the inherent meaning in us having an exchange as people,\u201d she says. To her, adding AI into relationships \u201cpresupposes inadequacy\u201d in them, and offers a sterile alternative: always saying the right thing, back and forth, frictionless forever.<\/p>\n The two women are no longer friends.<\/p>\n \u201cWhat you\u2019re giving away when you engage in too much convenience is your humanity, and it\u2019s creepy to me,\u201d Summer says.<\/p>\n Dr Mathieu Corteel is a philosopher and author of a book grappling with the implications of AI (only available in French<\/a>) as a game we have all entered without \u201cknowing the rules\u201d.<\/p>\n Corteel is not anti-AI, but believes that overreliance on it alienates us from our own judgment, and by extension, humanity \u2013 \u201cwhich is why I consider it as one of the most important philosophical problems we are facing right now\u201d, he says.<\/p>\n If a couple, for example, expressed love through AI-generated poems, they would be skipping crucial steps of meaning-making to create \u201ca combination of symbols\u201d absent of meaning, he says. You can interpret meaning retrospectively, reading intent into an AI\u2019s output, \u201cbut that\u2019s just an effect\u201d, he says.<\/p>\n \u201cAI is unable to give meaning to something because it\u2019s outside of the semantics produced by human beings, by human culture, by human interrelation, the social world,\u201d says Corteel.<\/p>\n