Thirty-one billion dollars. That is how much Americans spend annually on anti-anxiety medication. I sat with that figure for a while, turning it over, trying to feel its weight. Thirty-one billion is a number so large it stops meaning anything unless you put it next to something else. So here is something else. The cost of calling a friend is zero. The cost of eating dinner with your neighbor is the price of the meal. The cost of sitting on a park bench next to another human being is nothing at all. We are spending the GDP of a small nation to chemically manage a feeling that, in many cases, has a relational cause.

The Prescription That Cannot Be Prescribed

I want to be careful here because I am not anti-medication. Anxiety disorders are real neurological conditions, and pharmacological intervention saves lives. Dr. Kristin Neff's research at the University of Texas has shown that self-compassion practices significantly reduce anxiety symptoms, and she is the first to acknowledge that some people need medication to reach the baseline where self-compassion practice becomes possible. The pills are not the problem. The problem is that we have built a system where a person can see a psychiatrist, receive a prescription, fill it at a pharmacy, and complete that entire circuit without a single professional asking how many close relationships they have. Without anyone asking when they last felt genuinely known by another person. The Cigna 2024 Loneliness Index found that the majority of Americans who report high anxiety also report high loneliness. The overlap is not coincidental. It is diagnostic. But loneliness is not in the DSM, so it does not get a code, and what does not get a code does not get treated.

Following the Money Backward

Thirty-one billion dollars flows through pharmaceutical companies, pharmacy benefit managers, insurance networks, and retail pharmacies. Each of those entities has a financial interest in the continuation of that flow. Nobody has a financial interest in you calling your mother more often. Nobody profits when you and a coworker take a real lunch break together instead of eating at your desks in parallel silence. The economic incentives of American healthcare are structurally misaligned with the relational solutions that the evidence actually supports. The U.S. Surgeon General's 2023 advisory on the loneliness epidemic recommended community-based interventions, social infrastructure investment, and a fundamental rethinking of how we design public spaces. These recommendations cost relatively little compared to thirty-one billion annually. They also do not generate revenue for anyone. That is the tension. The cheap solution does not have a lobby. Dr. Julianne Holt-Lunstad's work at Brigham Young has demonstrated that strong social bonds reduce anxiety, lower blood pressure, improve immune function, and extend life. The effect sizes are not small. They are comparable to the effect sizes of the medications we spend billions on. But you cannot patent a friendship. You cannot mark up a conversation. You cannot run a direct-to-consumer ad campaign for the experience of being truly heard. I think about the person sitting in a pharmacy drive-through, waiting for a refill, scrolling through a phone full of contacts they never call. I think about what would happen if someone offered them a different kind of intervention. Not instead of the medication. Alongside it. An AI companion is not a replacement for human connection, but it is a space where you can practice vulnerability without the stakes that make vulnerability terrifying. Sometimes the reason people do not call a friend is that they have forgotten how. They have forgotten what it feels like to say something honest and have it received. Thirty-one billion dollars says we are desperate for relief. The question is whether we are willing to look for it somewhere other than a bottle.