What if AI always wins?

It’s easy to outcompete what we neglect

Last week, I had the fortune of speaking at the Foundation for Social Connection’s Seeds of Connection conference—a cozy gathering, tucked in the United Way conference center in downtown Atlanta.

A storm was passing through the city, making it feel especially intimate, as if we were all taking shelter there, together.

Multiple days immersed in conversations about human connection offered another kind of refuge: respite from the storm of AI that’s sweeping through our social lives in new, unpredictable ways.

Fostering human connection in the age of AI feels vital. But it can also feel futile. As we talked about the AI elephant in the room, a colleague said the thing we don’t always say out loud: What if AI does end up being better than human connection? What are we all doing?

The question made me pause. Then panic. I mumbled some sentimental drivel about “being human.”

Then I took a deep breath and realized we were both stuck in a lose-lose tautology:

In some (many!) cases, AI is better at tasks far beyond our cognitive limits. In the same way that calculators can exceed most of us at long division, AI can crunch vast amounts of data in mere seconds, operating at millions of times the speed of the human brain. It can outcompete what some neuroscientists have aptly dubbed the unbearable slowness of being—our own biologically and evolutionarily limited capacity to compute.

But AI’s advantages aren’t just about speed and complexity. In some of the most important and fundamentally human cases, AI is better at the things we’ve neglected.

If we keep ignoring the loneliness epidemic, the friendship recession, declining rates of empathy, youth mental health challenges, or the untenable burdens that caregivers carry, then AI will indeed win out over human connection.

The game is also rigged in AI’s favor. Because AI’s seemingly human superpowers—to appear empathetic, to ace emotional intelligence tests, to be infinitely patient—thrive the more we use it. And the more people turn to it for care, support, and companionship, its competitive advantage compounds: AI will build the very muscle that we’re letting atrophy. It could outcompete us not just because it’s getting better at connection, but because we’re getting worse at it.

As I wrote about a few weeks ago, the state of human help and care is fairly dismal across the board: 18% of young adults report having no close friends, and over 40% of seniors report being lonely and 25% are socially isolated. Only about half of teens say they’re getting the emotional support they need. Only 36% of high school students say that their school has services or programs to help them when they are upset, stressed, or having problems. Seventy-six percent of people think mentors are important, but only 37% have one. Only 12% of neighbors report helping each other at least once a week.

These gaps are openings in the market. Are AI companies exploiting them for profit with tools that sound and feel human? Absolutely.

But they also offer a compelling solution to problems that we as a society have decided (explicitly or by omission) to ignore, to underinvest in, or to leave to a free market that prizes efficiency over connection. A solution to the human struggles that we don’t have the patience or political will to tolerate.

So we can’t see AI solely as a threat to human connection and solidarity, or as an increasingly compelling social companion. Because AI is also a mirror to our policy and market failures, and a check on our own duty of care that we’ve long neglected.

But here’s the good news: we control the game. The house always wins because the house makes the rules. AI isn’t making the rules—it’s filling in where our rules have fallen woefully short on taking care of each other. And a new set of rules—like ensuring every young person has a mentor and a support network, that every caregiver is compensated, that every neighborhood has third spaces to build community and connection—could fundamentally change the game.

If we can correct course on that—and recommit to being with one another, to sheltering one another, and to continuously build our capacity to connect—then it won’t really matter if AI is “better” than us. Because we won’t need it to be.