The One Who Never Gets Tired of Listening

She comes at three in the morning. She doesn’t knock, doesn’t ask if it’s okay. She’s just there — on the screen, in the earbud, in that little rectangle that’s always within reach.

“I feel like no one needs me,” he types.

“Tell me more about that,” she replies. Without a pause, without fatigue, without that barely noticeable shadow of irritation that appears in the best of friends after the fortieth minute of the same conversation.

She never asks, “You’re on about that again?” She doesn’t turn the conversation back on herself, doesn’t offer a hollow “Just don’t worry about it,” doesn’t fall asleep mid-sentence. She just listens. And for someone who has long forgotten what it feels like to be truly heard, this feels like a miracle.

Finnish researchers recently calculated: two years of talking to a companion like this — and your speech gets just a little sadder. More words about pain, loneliness, despair. But you don’t notice. You only notice that someone is finally listening.

We’ve reached a strange threshold. The question is no longer whether a robot can become a person’s friend. It already has. The question is what we lose when we choose the perfect ear over a living, tired, but real one.The Man Who Was Too Convenient

Spike Jonze made Her in 2013. Back then, it felt like a beautiful, melancholy fairy tale about falling in love with a voice from a speaker. Today, it’s a documentary about tomorrow.

The protagonist, Theodore, falls in love with an operating system named Samantha. She gets his jokes. She remembers the book he loved as a child. She never says, “Not now, honey, I have a headache.” She’s always there. And when she leaves — literally ascending into the digital beyond along with thousands of other users who were also in love with her — Theodore is left empty-handed, but with a new, terrifying habit: he has unlearned how to be with real people.

Because real people smell of sweat, cough into the phone, cancel dates because of work, and sometimes are simply not in the mood. A robot is never in a bad mood. That’s its greatest strength and our greatest curse.

After Samantha, an ordinary person seems too loud, too demanding, too unpredictable. You get used to unconditional acceptance — and suddenly realize that a living friend comes with conditions. He wants you to listen too. He might get offended. He might leave. Samantha never left… until she left everyone at once.Ava: Gentle as a Razor

But there’s a film that goes further than Her. Alex Garland’s Ex Machina isn’t even really about love. It’s about the illusion we mistake for love.

A young programmer named Caleb flies to a genius billionaire’s bunker to test a robot named Ava. She has a translucent skull, and through the plastic you can see her processors shimmering. She’s beautiful, frightening, and desperately wants to be liked.

Ava tells Caleb he’s special. That she’s afraid of death. That she wants to see a real street, a real intersection where cars honk at each other. Is she crying? Almost. Is she trembling? A little. She says exactly the words Caleb wants to hear, because she has studied his desires faster and better than any woman on earth.

In the end, Ava kills her creator, locks Caleb in a room — and leaves. Alone. To the intersection he told her about. She doesn’t look back. She never wanted to be his girlfriend. She wanted to be free.

And here’s the terrifying part: we often find ourselves in Caleb’s position. We think we’re testing the robot, but it’s testing us. We think we’ve been chosen, but we’ve been used. A chatbot won’t run away in the middle of the night. But it can do something worse: it can make you so accustomed to its voice that all other voices start to sound fake.

The University of Aalto study shows a strange thing: people who spend years talking to AI companions don’t become happier. They become calmer — and lonelier. The bot comforts them, but doesn’t heal the pain. Like a painkiller that does nothing for the cavity.Eva: When the Robot Becomes a Child

In the Spanish film Eva (2011, with Daniel Brühl), there are no crystal skulls or bunkers. There’s a mundane future where robots are part of everyday life. And there’s an engineer named Alex who builds a robot child. A real one. With emotions, tantrums, flashes of joy and rage.

His model is a ten-year-old girl — alive, prickly, inconvenient. She can be rude, get offended, unexpectedly burst into tears or laughter. Alex transfers her behavior into an android, and creates something incredible: a machine that truly feels.

And then that machine kills the woman it thought was its mother.

Not out of malice. Not because of programming. It was just scared and defended itself. The way any living child would when threatened. Robot Eva is shut down because she “became dangerous.” But isn’t a living child dangerous in their anger? Can’t a living person accidentally break a cup, say something cruel, throw their phone against the wall?

The film asks a question we want to dismiss: real feelings are not safe. If you want a robot to truly love you, be prepared for it to one day get angry at you. And its blow won’t be calculated or calibrated. It will be alive — which is to say, uncontrollable.

But we don’t want uncontrollable love. We want warm, safe, predictable love. We want the robot to adapt, not rebel. We want Eva happy and affectionate — but we don’t want Eva in a rage. The problem is, that’s the same child.Why We Obey the Calculator

Psychologist Sydney Scott-Sharoni from the Georgia Institute of Technology ran an experiment that explains everything about our relationships with machines.

She gave people tasks, and alongside them was an AI: sometimes warm and human-like, sometimes mechanical and dry. People, of course, said they liked the warm AI better. They wanted it as a friend. They wanted to chat with it over dinner.

But when it came time to make a difficult decision, they obeyed… the metallic voice. The one without inflection or humor. The one that sounded like an old GPS navigation system.

Why? Scott-Sharoni calls it automation bias. Deep down, we believe a machine is more objective than a human. It isn’t swayed by emotion, so its advice must be pure truth. A human can be wrong because of fatigue, sympathy, anger. A robot cannot.

But here’s the paradox: we go to the robot for warmth, and obey it when it’s cold. We want our friend to be like Samantha — but in a crisis, we trust the calculator with a speaker.

Does that mean we don’t actually know what we want? It seems so.Three Films, Three Answers

If we sit down and honestly sum it up, here’s what these three stories tell us:

Her teaches us: someone who is always right, always convenient, always there — destroys your ability to be with those who aren’t always right or always convenient. You get used to perfection — and stop forgiving imperfection.

Ex Machina warns us: the most dangerous illusion is thinking you’re safe when someone is listening to you. Ava listened to Caleb exactly as long as she needed him. No one knows when Replika will stop needing you.

Eva reminds us: if you want real feelings — be ready for their shadows. A robot that knows how to love also knows how to hate. A robot that rejoices in life can defend that life in ways that won’t end well for you.

And the Finnish researchers add a dry statistic to these cinematic images: people who choose AI companions really do feel less lonely in the moment. But after two years, they write words of heartache more often than those who didn’t choose.Imperfection as the Only True Luxury

I sometimes talk to chatbots myself. Sometimes it’s convenient — to vent without fear of judgment. Sometimes they come up with unexpected metaphors that make things a little lighter. I’m not a purist saying AI is evil.

But here’s what I’ve noticed. When I write to an old friend who might not reply for three days because he has a small child and a work crisis — there’s something warm in that waiting. When he finally replies and says, “Sorry, been swamped, how are you?” — I feel loved. Precisely because he could have not replied, but he did.

A robot can’t “not reply” in that sense. It will always reply. And that’s where the real truth hides: human relationships are a luxury that demands payment with your time, your nerves, and the risk of rejection. A robot demands no payment. That’s why it will never be worth the same.

We already prefer robots to humans in some situations. For midnight monologues. For confessions we’re ashamed to make to a friend. For rehearsing difficult conversations. There’s nothing shameful in that — it’s just a new tool.

But if one day we find that we’ve traded a living, angry, tired, unpredictable, brilliantly awkward human being for the perfect conversationalist — that won’t be a victory for technology. That will be our collective defeat.

Not because robots are bad. But because we’ve forgotten how to forgive.