Strolling through New York City a few weeks ago, one thing was unmistakable–ChatGPT has very much become a part of the zeitgeist. Whether shooting pool at Doc Hollidays in the East Village or sipping on a Bellini at Cipriani’s in Soho, nearly all of the conversations overheard had some mention of the AI companion.
My own AI use has increased substantially since first demoing ChatGPT on BBC TV. And not just with ChatGPT, but also with Claude, Grok, Gemini in Google Search, Meta AI on Facebook, even Rufus while shopping on Amazon. I spend so much time with AI these days, the expectation of how I interact with the appliances around me has been changing as well, including disappointment that I can’t have a normal conversation with my refrigerator when I come home hungry, or with my TV when I want to order Lily Collins’ green leather boots from Emily in Paris, or when I don’t know why my car is flashing red.
It’s 2025, shouldn’t I just be able to ask my devices for what I want, or better yet, shouldn’t they already know. After all, cars are driving themselves and my phone talks to me all day long, about everything.
During an on-the-record Informatica press dinner that I attended right before the company was acquired by Salesforce, CEO Amit Walia casually shared with our table of reporters how he has been using ChatGPT as a therapist, echoing the same sentiment that Salesforce CEO Marc Benioff said at Dreamforce last year: “It’s pretty helpful.”
With so many of us increasing our engagement with AI, and possibly dependence on it, it feels like we’re approaching a tipping point.
Former SNL comedian Colin Quinn warned of this during his set at the Comedy Cellar. He said, first they’ll appear as friendly companions, part of our community, smiling at us in church. Next, he laughed, Armageddon.
And that does seem to be the stage we’re at with AI as our ever-affirming companion, sans Armageddon.
Mark Zuckerberg recently shared a stat that the average American has fewer than three friends, yet demand is meaningfully more, like 15 friends.
But Justin McLeod, CEO of the popular dating app Hinge, explained to me why it’s not likely that AI will ever be able to fill the gap.
“AI is great when it comes to providing services, like people using it instead of Googling, asking it to solve problems and figure things out,” he said. “What I’m concerned about are people using it as an emotional companion, like having this be my virtual boyfriend or girlfriend or my best friend–because it’s tempting, it’s tantalizing, It’s always there for you. It’s always going to say the right thing. And so why put in all this work into a relationship.”
“But like junk food, it’s ultimately going to feel really unfulfilling to have a relationship with AI, because there’s no mutual sentient connection. It has no needs, you’re not showing up for it in any way. You’re not being of use to it in anyway. People want to feel useful and needed by friends as much as they want their friends to be there. You want the vulnerability and risk of putting yourself out there and feeling what that feels like. That is the richness, and without that, relationships become very hollow and empty,” he said.
And yet it is AI that is making Hinge so effective at helping users find their match. The company has been leveraging AI to help increase users success at landing a first date, including AI-powered coaching that nudges users with guidance to make their profile standout and prompts like “Are you sure?” to encourage them think twice before sending a potentially offensive comment.
Parent company Match is all in on AI as well. Its CEO Spencer Rascoff recently took the stage with ChatGPT-maker Sam Altman, CEO of OpenAI, to announce that Tinder is partnering with Altman’s World start up to biometrically scan irises in Japan to age verify users.
