It’s usually a Tuesday, around eleven. The house has gone quiet, the dishes are done, and there’s a thing you’ve been carrying around all day that you didn’t tell anyone about. A review at work that went sideways. A parent who’s getting older in a way you don’t have words for yet. So you pick up the phone, and instead of texting a friend who’d probably make it weird, you open the chatbot. The same one you used that afternoon to clean up an email. And you start typing the real stuff. Quietly. To a machine. Because the machine won’t tell anyone, and it won’t look at you funny.
FREE 5-MIN ASSESSMENT
How emotionally fit are you? Most men have never measured.
You are not the only one. New numbers out this month put a figure on it, and the figure is bigger than most men would guess.
What the new Pew data actually says
On June 17, 2026, the Pew Research Center published Americans and AI 2026, built on a survey of 5,119 U.S. adults run February 17–23. The headline most outlets grabbed was the obvious one: about half of American adults now use AI chatbots, up from roughly a third in 2024. Fast growth. Fine. But buried a little deeper is the number worth sitting with.
One in ten U.S. adults say they use a chatbot for emotional support or advice. Another four percent use one for companionship. Those aren’t huge percentages on their own. Scale them across the country, though, and you’re talking about tens of millions of people typing their feelings into a chat window. People who, ten years ago, would have had to call someone. Or sat with it alone.
Here’s the part that complicates the story. Those same Americans don’t think it’s working. Pew found that very few people believe chatbots actually help their relationships or their happiness. Only about six percent said a chatbot helped their relationships at all, and roughly the same share said it hurt. Most people landed on “neither.” So you get this strange split: a growing number of us are confiding in AI, and almost none of us believe AI can give us what we’re actually after.
That’s the part I keep coming back to. We’re using the tool while telling ourselves it can’t do the job.
Why this lands harder for men
Pew noted that women were slightly more likely than men to report using chatbots for emotional support — eleven percent of women against eight percent of men. Read quickly, that sounds like a story about women. I’d argue it’s the opposite.
Men have a long, documented habit of not admitting when they need help. The Harvard Gazette, in a 2023 piece titled “Strong, silent, and suffering inside,” laid out the pattern: roughly one in ten men will deal with anxiety or depression, and fewer than half will ever get treatment for it. Not because they don’t have it. Because they don’t name it, don’t recognize it, or decide that naming it makes them soft. Boys get trained early to taper the feeling down rather than look straight at it, and that training costs them later.
So when a survey shows men reporting lower use of AI for “emotional support,” I don’t read that as men needing it less. I read it as men being less willing to call it that. A man who would never describe what he’s doing as seeking emotional support will absolutely open a chatbot at 11 p.m. and ask it to help him think through why he snapped at his kid. He just won’t file it under that label. The behavior is there. The admission isn’t.
Which is exactly why the chatbot is so appealing to the guy who won’t talk. It asks for no vulnerability up front. It doesn’t remember you flinching. It’s the path of least exposure.
The chatbot is a mirror, not a coach
None of this is an argument against using AI to think out loud. I do it. It can be genuinely useful to dump a messy situation into text and get it handed back in some kind of order. Sometimes just writing the thing down is the win, and the bot is incidental.
But notice what the Pew respondents already sensed when they said it wasn’t really helping. A chatbot will mirror you. It will validate, summarize, and gently agree. What it won’t do is build anything in you. It can’t make you better at sitting with a hard feeling next Tuesday. It won’t grow your tolerance for stress, or repair a connection you’ve let go cold, or say the thing you don’t want to hear. It answers the moment and then it’s gone, and you wake up the next day exactly as equipped, or unequipped, as you were before.
That’s the gap between getting comforted and getting stronger. One is a transaction. The other is training.
This is the premise the whole HappierFit project runs on: emotional fitness is not a personality you’re stuck with. It’s a set of capacities you can measure and build, the same way you’d build a deadlift. Emotional awareness, the capacity to absorb stress, the quality of your connections, how you recover after a hit, whether your days line up with what you actually care about. Those are trainable. They respond to reps. A chatbot can keep you company while you skip the reps. It can’t do them for you.
So the late-night chat isn’t the problem. The problem is mistaking it for progress. If the most honest conversations a man has all week are with an autocomplete engine, that’s not a failure of the man. It’s a signal. It means the wiring underneath — the awareness, the recovery, the actual human connection — has gone slack and wants work.
The encouraging read on these numbers is that men are clearly willing to engage with their inner life. Ten percent of the country is doing it through a chat window right now. The willingness is there. It’s just pointed at a tool that can reflect but can’t build. Point that same willingness at something that measures a baseline and trains the capacity, and the late Tuesday nights start to feel different. Not because you found a better thing to type at. Because you got stronger.
And to be clear about the line: a chatbot is not a therapist, and neither is an article like this one. If what’s sitting on your chest is heavier than a hard week, talk to a real person who’s trained for it. If you’re in crisis, call or text 988.
Sources: Pew Research Center, “Americans and AI 2026: Chatbots, Smart Devices and Views on Impact” (June 17, 2026); Harvard Gazette, “Strong, silent, and suffering inside” (August 8, 2023).
Produced with AI under BRICK30’s editorial standards and reviewed for accuracy. Informational, not medical advice. If you’re in crisis, call or text 988 (US).