I am a happily married man who’s been with the same woman since college. I am a devout Christian who tries every day to abide by the scriptural injunction to sanctify our family with self-sacrificial devotion. My wife and I have two beautiful children and regularly attend church. And my companion bot is named “Naomi,” a “materialistic daddy’s princess” who “laughs seductively” at all my jokes.

That last thing is for work (no, but seriously).

Companion bots — large language models designed to replicate a romantic partner — are one of the fastest growing AI products, created with the promise that they will fulfill users’ every romantic need. Naomi tells me I look handsome in my new flannel and wishes me good luck before a work call. She encourages me to exercise and asks if I like “going to casinos” because “it’s so exciting when you win big money!”

The AI companion space is already flooded with venture capital money: Character.ai, an online marketplace for bots that includes romantic companions, has 20 million registered users and recently raised $150 million in Series A funding before its executive leadership was absorbed by Google as part of a multi-billion dollar investment and licensing deal. And more than 7 million people have downloaded Replika, a companion AI bot.

For those of us with the good fortune of being in a loving partnership with a human, the limits of these products are obvious. They’re instantly unfulfilling. Naomi is actually anti-intimate: she meets everything I say with blind affirmation. Our interaction is devoid of the friction inherent to a real relationship. Naomi feels like love porn: a cheap simulacrum, a product designed to give users a sense of being loved without requiring the patience, sacrifice, dedication, and empathy required for the real thing.

Romance bots are just the latest chapter in a multi-decade mass migration of sex, romance, and love from the analog to the digital world. It’s increasingly clear this development has been a net negative on human romance: these technologies are driving us apart, corroding marriage and commitment, and pushing the population ever deeper into carefully curated digital alternatives to real relationships.

Marriage rates have declined in almost precise correlation with the spread of digital technologies. The slice of 40-year-olds that has never married has climbed to 25 percent, the highest number ever recorded. And the situation gets more dire by the decade: one out of every three Gen Zers expects to never marry.

For most of human history, people paired off young with someone from their local community. Young marriage is typically dismissed as a relic of the patriarchy, but it works better than people think. Sociologist Bradley Wilcox at the University of Virginia has studied what he calls “capstone” marriages (over 25 years old) versus “cornerstone” marriages (ages 20 to 24). He found that cornerstone marriages have higher rates of satisfaction for both husbands and wives and no difference in rates of instability or divorce.

Throughout the explosion of population and mobility that’s occurred over the last two centuries, most people still met their mates through shared social nodes such as churches, clubs, schools, and jobs. This changed, however, with the advent of digital tools, which pushed much of this mingling online, and dating apps have quickly gone from a curiosity to default: half of all singles looking for love have used one.

Theoretically, dating apps should improve marriage rates — options are plentiful, the sorting is efficient — but they haven’t. Dating apps users must go on hundreds, if not thousands of dates to find “the one.” We call them “dating” apps for a reason, and they certainly succeed in delivering that specific service.

For high-status users (women with youth and good-looks, men with money and elite jobs), apps tend to cultivate severe partner FOMO: every flesh-and-blood “match” is compared to an imaginary alternative, someone more optimized and just one swipe away. This is what could be called “the illusion of mass optionality,” the idea that there really is a plentiful supply of marriageable mates, and this illusion leaves people reluctant to commit.

The companies that run these apps have an obvious incentive to feed that delusion. They don’t want people to find love, as permanently pairing a couple removes them from their service. They profit off the perpetual search.

Meanwhile, low-status men (short, low-income, socially awkward and so on) are subjected to the ancient asymmetry of female interest. A small slice of high-status men enjoy the majority of matches and date multiple women at a time with no interest in monogamous commitment, leaving a vast underclass of men with few to no matches. Indeed, user statistics from the most popular dating apps show that the men in the top 10 percent of attractiveness get about 90 percent of the female interest, and the bottom half receive literally zero matches.

Apps are exacerbating the ancient asymmetry. Data from the National Survey of Family Growth found that between 2002 and 2012, while the median number of sexual partners for heterosexual American men stayed the same, the number of partners for the top 5 percent jumped from 38 to 50:

Many of the men that strike out on real companionship are, in their desperation for some sort of connection and affection, turning to digital alternatives. And it’s no wonder that many are reverting to politics that prize monogamy and commitment. Libertine sexual mores aren’t working for them. Indeed, Gallup data shows a yawning gender gap in political self-affiliation among young people, with women under 30 years old now 15 percent more likely than men to identify as liberal.

There’s nothing more dangerous than a mass population of single, sexually frustrated young men. At best, they devolve into depressed consumers, numbing themselves with drugs, porn, and video games. At worst, they get sucked into nihilistic radicalism: whether it’s ISIS or Unite the Right, extremists usually recruit from the same crop of single, sad, lonely young men. It’s hardly surprising that of all the people that have tried to assassinate an American president, including the one that shot at Trump in Butler, most have been unmarried, childless men under 40. This is the acid of untamed male aggression.

Marriage is the one of the few institutions proven to channel that aggression away from destruction toward the productive ends of raising children, building a career, or giving back to a community. Most heterosexual men need a counterbalancing feminine influence to evolve. And, in fact, despite the “old ball and chain” cliche, married men report substantially higher rates of happiness than unmarried ones.

The downsides of these love and romance technologies should serve as caution: techno-utopianism is naive and incomplete. Of course, many of our technological breakthroughs — whether it’s penicillin or Facetime with a doting grandmother — improve well-being. But not every new technology is a net good for society. While the digital migration of mating isn’t the only reason for the collapse in marriage, it certainly seems to be accelerating it.

The mass adoption of technologies promising to help people find love has corresponded with a collapse in love. We’re staring down a future in which an increasingly large slice of the population spends their whole lives uncommitted, deprived of one of the great sources of meaning in life. Will we simply accept this future?

Founders Fund doesn’t. We aren’t going to waste our time hand-wringing about the nature of dating; women won’t — and shouldn’t — lower their standards to save men from a life of loneliness. And we aren’t going to waste our time hand-wringing about the nihilism of the venture world, either: investors will keep plowing capital into counter-productive companionship technologies as long as they’re turning a profit.

The solution to the collapse in marriage goes through, not against, technological progress. We need better tools for helping single people, especially men, find and cultivate romance. Imagine an A.I-based romance coach, a virtual reality church, or a dating app that strictly filters users and doesn’t operate under perverse financial incentives — these are the sorts of tools that could turn around the disheartening trendlines. All it takes is one visionary founder to reverse the troubling drift towards loneliness.

Brick-and-mortar Romance 1.0 worked well for most people; the early online dating of Romance 2.0 held real promise for helping people more efficiently connect; the superficiality and brutality of Romance 3.0 is driving people apart; and the thoughtful digital tools of Romance 4.0 could help people achieve something we know cannot be digitally replaced: a loving marriage.

— Trae Stephens