Researchers at the University of British Columbia and the University of Pennsylvania split 296 first-semester college students into three groups for two weeks. One group texted a randomly assigned fellow student every day. One chatted daily with “Sam,” a chatbot the researchers designed to be the ideal supportive friend. The third group wrote a single sentence in a journal.

Sam won on empathy. The researchers’ analysis of the chat logs showed the bot expressed more care and understanding than the human partners did. Sam also lifted the students’ mood more than journaling did.

But when the researchers measured loneliness, only the students who texted another person felt less lonely. The Sam group ended up no better off than the journal group.

Why? Lead researcher Ruo-Ning Li offered this: “When you’re talking with a chatbot, you can get a lot from it, but you never have the chance to give something back.” The data fit her point. Students showed less empathy toward the bot than they showed toward their human partners.

As a former electrical engineer, I’d put it this way: the chatbot delivered signal but never closed the loop. Human connection is a feedback circuit. Care goes out, care comes back, and each strengthens the other. Remove one direction and nothing circulates.

This lines up with something I wrote in Seeking Sanity: loneliness arises from the mind’s belief in its isolation. Receiving care soothes that belief. Giving care dismantles it, because it’s hard to feel like a small, fragile, separate self while you’re actively tending to someone else. That’s why loving-kindness and tonglen, both practices of giving, work so well. It’s also what I mean by being “wisely selfish”: caring about others’ needs as much as your own.

Now, let’s not overreach. Sam didn’t harm anyone, and easing a bad mood has real value. I think of a chatbot as a painkiller: useful, but not a cure. The researchers also point out that college students have plenty of chances to socialize, so a chatbot might help more for people who are truly isolated, such as elderly adults. One two-week study of freshmen doesn’t settle anything, and I’d be a hypocrite to claim it does.

So what do you do with this? Dr. Li’s advice is wonderfully low-tech: reach out to a person, whether a classmate, a neighbor, or the barista. I’d add one twist. This week, send one message a day whose only purpose is to give: a thank-you, a check-in, a “thinking of you.” Ask for nothing back. See if anything shifts.

Your turn: Who could use a message from you today?