AI Wants to Be Your Matchmaker. It’s Going to Ruin Love.

You’ve felt it. That dead-eyed scroll through Tinder at 11pm, thumb moving on autopilot, swiping left on faces that blur into one composite human. You’re not looking for a person anymore — you’re looking for a feeling, and the app has made damn sure you’ll never find it.

Now the tech industry has a solution: AI. And the very people who built the dating apps that broke modern romance are the ones now promising to fix it.

Justin McLeod, the founder of Hinge — the app literally branded as “designed to be deleted” — just quit his own company to launch Overtone, an AI-driven dating startup that just raised $18 million in seed funding from investors including Match Group. Yes, Match Group: the company that owns Tinder, the app that arguably created the swipe-fatigue crisis in the first place.

In China, a former Kimi AI search tech lead launched 良配AI (Liangpei AI), backed by Capital Today — the same firm that bet on JD.com, Meituan, and BOSS Zhipin. Their pitch? “Get married or get a refund.”

And then there’s DUO, a U.S. startup that doesn’t even bother building an app. You just text an “AI matchmaker” in iMessage, have one deep conversation, and wait. The AI evaluates compatibility across “hundreds of dimensions” and pushes matches to you directly. No profiles. No swiping. No app.

It sounds like progress. It feels like relief. It’s neither.

The fundamental problem with AI matchmaking isn’t that it won’t work. It’s that it will work — and in working, it will quietly kill the thing that makes love worth finding.

Here’s what the founders get right: the current dating app model is broken. Swiping through photos and reading “love hiking and tacos” bios was never going to reveal whether someone shares your values, your communication style, or your vision of a good life. The swipe economy optimized for engagement, not connection. It gave you more options and fewer relationships. Everyone knows this. The founders know this too — that’s why they’re pivoting.

But the AI fix assumes a dangerous premise: that love is a compatibility problem to be solved with better data.

It’s not.

Love is irrational. It’s the person who doesn’t match your “type” but makes you laugh in a way no one else can. It’s the boring-on-paper partner who turns out to be the most interesting person you’ve ever met. It’s the friction, the surprise, the moment of serendipity that no algorithm can predict because it doesn’t follow a pattern — it breaks one.

When you optimize for compatibility, you don’t get better matches. You get a filter bubble for the heart.

Think about what happens when an AI evaluates “hundreds of dimensions” to find your perfect match. It’s doing what every recommendation algorithm does: narrowing your world to what it predicts you’ll like. Netflix does this with movies. Amazon does this with products. Now AI wants to do it with people.

But here’s what recommendation engines never tell you: they don’t show you what you need. They show you more of what you already are. They reinforce your existing preferences instead of challenging them. They turn you into a narrower version of yourself.

Apply that to love, and you get relationships that feel curated rather than discovered. You get partners who fit your data profile but not your soul. You get the romantic equivalent of an echo chamber — two people who are “compatible” in every measurable dimension and bored to death by each other within six months.

And there’s a darker layer. Who decides what “compatibility” means? What values get weighted highest? What happens when the AI decides that political differences reduce compatibility scores, or that income gaps matter, or that certain personality types “shouldn’t” match? Every algorithm carries the biases of its creators, and when the product is human relationships, those biases don’t just shape content feeds — they shape lives.

The DUO model is especially revealing. One conversation with an AI, and it decides who you should meet. No browsing. No agency. No serendipity. It’s efficient, sure. It’s also the romantic equivalent of being assigned a roommate in college — except the assignment is based on a chatbot’s interpretation of one conversation, and the stakes are your emotional future.

The founders will tell you AI understands nuance now. It can read between the lines. It can detect communication patterns and emotional needs that a questionnaire never could. And that’s true — partially. Large language models are genuinely better at parsing conversational subtext than any tag-based system that came before.

But understanding what someone says they want is not the same as understanding what they need. Humans are famously terrible at articulating their own desires. We want excitement but choose stability. We say we want honesty but fall for charm. We describe our ideal partner and then fall in love with someone who matches none of it.

An AI that perfectly understands what you say you want will perfectly deliver the wrong person.

This is the paradox at the heart of AI dating: the better the algorithm gets at matching on measurable dimensions, the more it squeezes out the unmeasurable dimensions — the ones that actually matter. Mystery. Tension. Growth through difference. The discomfort of being with someone who challenges you in ways you didn’t know you needed.

None of this is in the training data. None of this survives optimization.

And let’s be honest about what’s really driving this trend. It’s not just about helping people find love. Match Group investing in Overtone isn’t charity — it’s a business bet. Dating apps have a fundamental business problem: when they work, users leave. “Designed to be deleted” is a beautiful slogan and a terrible business model. AI matchmaking promises higher match quality, longer retention, and — most importantly — higher lifetime value per user. The incentive isn’t to help you find love and leave. The incentive is to keep you engaged with the matching process itself.

The business of AI dating isn’t finding your soulmate. It’s making the search feel productive enough that you never stop paying for it.

This isn’t to say AI has no role in dating. Used thoughtfully, it could reduce the exhausting friction of modern dating apps — the endless profile-scrolling, the ghosting, the dead-end conversations. It could help people articulate what they’re actually looking for. It could surface connections that proximity and social circles never would.

But that requires a fundamentally different design philosophy: AI as a tool that expands possibility, not one that narrows it. AI that introduces you to someone you’d never swipe right on. AI that optimizes for surprise, not predictability. AI that understands the difference between a good match on paper and a real connection in life.

Right now, every AI dating startup is building the opposite. They’re building systems that promise to know what you want better than you do. And maybe they will. But the day an algorithm knows what your heart wants better than you do is the day your heart stops being yours.

So yes, the swipe era is ending. The AI matchmaking era is beginning. And somewhere between the two, we need to decide whether we want love to be something we discover — or something we’re assigned.

Because here’s the truth nobody in Silicon Valley will say out loud: love was never supposed to be efficient. It was supposed to be a mess. And the mess is the point.