# AI as Emotional Infrastructure for Relationship Intelligence

**Podcast:** FT Tech Tonic
**Published:** 2026-10-14

## Transcript

Hello, I'm Madhumita Merjia, the FT's AI editor.
Tectonic will be back soon with brand new episodes.
But in the meantime, we want to share some of the best bits from the FT Weekend Festival held in London last month.
At the festival, I got the chance to speak to the writer and philosopher Alain de Botton.
He has written about relationships, love, and the pursuit of happiness for many years.
I wanted to talk to him about artificial intelligence.
and the impact that human-like chatbots might have on our emotional lives.
So here's the conversation recorded at a packed tent at the festival.
I hope you enjoy it.
Hello, hello.
Thank you all for packing up this tent to come and see me.
So nice to see you all.
Alain, I heard we have tried to get you here for many years, and so I'm so pleased that you've decided to sit down with us here today.
I'm, you know, completely, let me just say, out of my depth when it comes to your expertise, I have spent 12 years writing about AI.
And I think I'm here just as much as everyone else to ask you to be an oracle for the thing that no matter who you are or how old you are, we all want to know, which is...
what is going to happen to our relationships, what happens to us as humans and the way that we connect to one another as this technology shapes and changes the way that we all live and work and love.
So hopefully you can tell us some oracle truths by the end of today.
You need no introduction, but Alain is a writer, an essayist and the founder of the School of Life.
I was looking back at all your work and I thought it was really fascinating that you've been writing about love since I think you were 23 years old.
So I wondered if you could start by telling us what young Alain got right and wrong and what you have learned, what you think now is different to what you were writing when you were in your 20s.
Such a good and humbling question.
Thank you so much.
I mean, as for the oracle bit, obviously I don't know much about...
But what I do know about is human nature.
And I think that some of the ways in which we can guess where technology is going is to think at ever greater depth about human nature.
Because, you know, you don't need to know too much about machines.
You need to know about the gap in which machines may enter in order to read some of the future.
But to come back to your question about the younger me and what one learns with age, I mean, it's a frightening question.
I mean, I think that the younger me would be surprised at how difficult relationships...
have continued to be.
I was very puzzled by love at 23.
I'm still puzzled by love now at 56.
I think that the great discovery for me is the unconscious.
That could sound quite strange, but let me try and explain.
I think one of the weirdest things about the way in which we operate is that...
Only a small part of our minds is actually open to conscious interpretation and interrogation, at least directly so.
A lot of what we do and feel is in the shadows, and we don't exactly know why we feel and act as we do.
It's one of the great sort of mysteries.
And psychoanalysts and psychotherapists describe this zone as the unconscious.
And the leading idea here is that understanding ourselves...
creates resistance in ourselves because so much of what we are is frightening.
We are made anxious by too direct a confrontation with a lot of what's actually inside us.
And so what we're often doing is being in flight from ourselves.
And there's a great cost to that.
The whole of psychotherapy is really based on the idea that...
running away from yourself is not a productive activity.
And that by standing still and feeling what you should feel, thinking what you should think, there will be benefits.
One view is that insomnia that some of us suffer from, many of us suffer from, is that it's really the mind's revenge for all the thoughts that you haven't had in the day coming to chase you at night, making sure that they have a chance to be aired.
And so I'm answering your question in a very meandering way, but I think I've come to to study really and to think about how much what we do in relationships is down to unconscious patterns, unconscious influences.
We think we're far more in charge than I think we actually are.
We don't really know why we're attracted to the people that we're attracted to.
One very dark thesis, a leading thesis in psychotherapy, is that really what we search for in partners is not so much people who will make us happy, but people who will, in one way or another, feel familiar.
And familiarity is a tricky one because what we may be most familiar with from our own past are patterns of suffering and unfulfillment.
We may feel that the The most familiar-feeling person is someone who doesn't take us seriously, belittles us, humiliates us, has something more interesting to do.
And, you know, it creates a great sort of paradox that sometimes people try and set us up on a date and say, you know, this person's perfect.
On paper, they're great, et cetera.
And we meet them and we do acknowledge that they're fantastic.
But somehow something holds us back from deepening our relationship to them.
And we might put it this way.
We might think, well, ultimately what's wrong with this person is that they're unlikely to make me suffer in the way that I need to suffer.
in order to feel that I am properly loved.
And so it's very, very paradoxical.
So I think I've been more and more alert to this strangeness.
And we have to treat ourselves with compassion and a dose of humor because we do get up to very odd things in love.
So despite that dark thesis, you have written that, I think you've said about 70 to 80% of our happiness is sort of down to the quality of our relationships.
Given that, you know, Just coming to the topic that I think about a lot, and I think you have recently written as well, what's the gap there that AI is coming in to fill?
Can you kind of lay out what is it doing here?
Why are we talking about AI in the context of something so human and also so...
I mean, look, the way I look at AI, I think there are huge continuities.
We know all about the disruptions that AI presents, and obviously it does in many areas.
But there are also huge continuities.
The project of AI is continuous with...
really the most ancient philosophical project, which is to bring intelligence to bear on the areas in which it's most needed.
Plato would not have been shocked by the ambition of those who lead AI.
And in fact, the very idea of a book, you know, there's often this sort of conflict between AI and books.
One way to think about AI is that it is...
It's the stuff that's in books delivered at a moment of particularly acute need in ways that are no longer constrained by the requirements of the press, of paper and ink and glue.
So in a way, we've always needed intelligence.
And maybe what your question is really pointing at is...
What's the kind of intelligence that we need in relationships?
And I think what we most need is what you might call perspective.
We are constantly in danger of losing perspective when it comes to relationships because we simply misread situations.
There's an incredible amount of bias from our past, from our history, from our emotional functioning, which means that we misread situations.
It's interesting.
I was talking to a friend the other day.
A lot of the criticism of AI is that it's sycophantic.
You hear this a lot.
is too flattering in its responses to us.
I look at this slightly differently in the context of relationships, because I think most people who turn to AI in moods of emotional distress around relationships, and a huge number of people do, these people are on the whole not suffering from arrogance, not suffering from hubris, not suffering from kind of egomania or feelings of omnipotence.
Insofar as they are feeling biased, they tend to be feeling biased in excessive humility, excessive deprecation, an excessive sense that they may be to blame for something in which they may actually be quite innocent.
I mean, that's just my sense of the people who turn to AI and indeed therapy.
I mean, I'm practicing and trained as a psychotherapist as well.
And so I see some of this from that end of things.
And I don't think the model should, I don't think we should be too scared of excessive sycophancy because as I say, very often when people are turning to AI, they're saying, I'm responsible for the entire collapse of my relationship, or why is everything I do wrong, or why will I always be unable to find someone, et cetera.
So it's a kind of excessive punitiveness that can weigh on us and that needs a corrective.
So really, we're talking about a corrective.
You know, a lot is said about AI and originality.
The great thing about AI in many areas, particularly in relationships, is that it's not particularly original.
And that's fantastic.
We don't need originality in this area.
What we need is, you know, the voice of the village elder.
We need common sense, as it were.
We need glorified common sense or deepened common sense.
And that's what AI is very good at doing.
It is the wisdom of crowds.
It's providing us with an aggregate common sense view of many dilemmas.
And when it comes to relationships, that can be extremely helpful.
That is precisely what we need.
We don't need anything more esoteric.
If you're writing a poem, maybe you need something totally original.
If you're actually saying, you know, what should I do now that I've been left?
Well, you know, collective wisdom may be exactly what you're after.
Okay, let's talk more about this because I completely disagree that it's going to fill that hole and I want you to try and convince me why.
But can you talk about your vision?
You've laid this out in an essay that you wrote that was published last month, I think it was.
about how you think AI will play a role or is playing already.
And you touched on this.
It's been one of the most surprising uses of AI and possibly one of the most popular ways in which chatbots have been used by people of all kinds, from young people are using it, but also age and gender independently.
And so tell me about how you see AI being used.
What does that mean?
What does a household look like where a couple is using AI in their daily lives?
Yeah, yeah.
So, look, I think that, I mean, it's a humbling thing because, you know, it's a kind of dethronement of the human being.
We're being asked to think that our brains, in the way we utilize them day to day, may not be particularly good organs for understanding and interpreting the conduct of our relationships.
And, I mean...
What we have to say is we're not doing very well unaided.
I mean, you know, things were not terrific before AI came along.
50%, as we know, of marriages end in divorce.
And of the other 50%, you know, well, I mean, the sun is shining.
We don't want to depress you.
But many of those will not be going great either.
We can say that, you know, pain is general in this area.
I would say that, you know, there's probably only...
10, 15% of people who can claim to be fully content in the area of relationships.
The rest of us are, in one way or another, suffering.
So it's really extremely problematic.
To try and get on with another human being is very, very difficult.
One of the things that we fail to understand is that there are two separate realities.
Whenever a couple come together, what you're really getting is, imagine, you can draw an analogy with a dictionary.
We've all got dictionaries in our minds, and those dictionaries have got entries appended to a huge variety of concepts.
So it'll be an entry, for example, on what a messy bathroom means, what being 10 minutes late means, what leaving the peanuts out in the kitchen means.
And each of us will have an entry on that.
And the thing is that we're very good at knowing our own dictionaries and not so good at checking in on the dictionaries of others.
And that's where conflicts arise.
So frequently you get couples fighting to the death.
Over something where they think they've got a disagreement because they think that they sit they're seeing the same entry for the same thing They may actually be in agreement.
They're just fighting over something completely different So, you know, it may be that for one person, a messy bathroom is a harbinger of an authoritarian and dictatorial style of living, whereas for somebody else, a messy bathroom is a harbinger of chaos and a truly unfair departure from order and good sense.
So we're extremely bad, really, at realizing that other people...
have different minds to ourselves.
We constantly project our own reality into the reality of others.
You know, this is why we get people who sulk.
You know, the sulk is a fascinating phenomenon that AI can help us with.
The basic idea behind a sulk is that somebody else should be able to read our mind for us, and that if they don't, they're evil, and they're a bad person.
And so...
So, you know, small children have this extraordinary experience.
When you're a baby, your parents do actually mind read.
They look at you and you've got a need.
You can't express the need in words.
You let out an inchoate scream and your parents will interpret it.
And, you know, on the whole, they might get it right.
But as we grow up, we continue to have this idea that if someone cares about us, they should intuit what's going on inside us.
And this is deeply problematic because our needs become ever more complicated.
So you get these...
tragicomic scenarios where, you know, one person upsets another, the person is too hurt to explain why they've been hurt.
So you cross your arms and the partner says, not being too autobiographical, the partner says, is something wrong?
And you go, no, no, nothing's wrong.
And the other one goes, well, you seem a little upset.
No, I'm not upset at all.
And then, you know, you go home and you lock yourself in the bathroom and the person's knocking, going, look, come out.
And you go, no, I don't want to.
Because essentially, if you put words to your mood, you're saying, if you loved me properly, you would know what was going on inside me without me speaking.
And the fact that I need to use words means that you're bad.
Now, AI...
AI could really help us to increase the data.
I mean, if you look at the future of medical treatment, it will be continuous monitoring of a human being.
So rather than checking in with a doctor once a year and catching things too late, you will catch them at the point of inception.
And something similar, it's very easy to imagine, is going to happen around relationships.
I mean, arguments are fascinating.
You know, very often arguments rumble over days, weeks, months.
At the end of a long period of sulking, of disagreement, of irritability, someone will go, you know, what is it really?
What is it?
And someone will go, well, three weeks ago, when I said that about my mother-in-law, you didn't laugh.
And you think, what?
And you've missed it.
You've entirely missed it.
And that has led to a retaliatory...
argument, a first retaliated argument involving granola, which then involved another retaliatory argument involving their cousin, and then, you know, and then you're on and on and on and on.
By this time, of course, you're not having sex.
How can you have sex with someone that you're furious with?
Most of the impasses in sex are not to do with a failure of attraction.
They're to do with fury.
You're furious.
You're not going to let yourself be touched by someone you're angry with, but you don't know what's made them angry.
And the person, the person may have lost touch with their own.
anger and may not be able to explain it to another person.
So things are constantly missed.
We forget what we're feeling and why we're feeling it.
AI, we may decide, I wouldn't mind, giving, you know, with the right safety protocols, giving some data over to an AI in order to monitor what on earth may be going on.
It'd be wonderful if someone said, you know, it's been two days since you checked in on that point of issue.
If you said, you know, You know, I'm still thinking about that.
That will hugely improve the chances of the weekend going well, et cetera, et cetera.
So we're really, we miss so much and it matters so much.
As I say, let's keep that figure in mind of the sort of 60, 70, 80% of life satisfaction dependent on a flourishing relationship.
The key to a flourishing relationship, they always tell us, is communication, right?
That the ability to feel heard is amazing, right?
If somebody says, I may not agree with you, but I hear your point of view.
I may not do exactly what you want, but I can see why you want it so much.
This is a magic salve.
I mean, it always works.
But we don't do it.
And we don't do it because partly out of retaliation, because the other person didn't do it, etc.
Imagine a world in which we are given many greater clues what the other person wants to hear from us and what we have missed.
This will hugely improve humanity.
Let's put aside the question of privacy and surveillance and letting the big tech companies know the deep insides of how we feel.
That's a separate thing.
But even on the question of is this going to solve the problems that you've laid out, do you believe that AI is a long-term solution?
You talk about human flourishing because what you're laying out sounds to me like a sort of short-term, this is going to fix.
this fight in this moment.
But a lot of what you've written before and what you've pointed to is how we learn to be better at loving or having relationships.
And part of that is the suffering that you've discussed and learning from that, right?
The friction.
Can we ever...
I don't know, have a relationship where AI is in our ear telling us to do something and actually learn from that?
And how do you see that leading to a long-term happen?
I mean, look, let's remember that the thing we call therapy has the shape it has for mostly out of sort of technological limitations.
I mean, why are sessions 50 minutes?
Why do people go once a week, et cetera?
These are to do with the sort of frictions of how...
The institution is organized.
But there's nothing, I mean, many of us know that the time you really need a therapist is not on a Tuesday at 5 p.m., but at 3 a.m.
on a Saturday night when no one's working.
That's when the problem is acute.
And we've always had these things called books, which you can turn to at a certain moment.
And we know that a book may have...
a really interesting and important thing to say.
It's just so devilishly hard to find it.
So we know that at certain points, you've got a particular argument or dissent on your hands in your domestic life, and you think, I know that someone somewhere has got an idea that would help.
Where is it?
And we can't find it.
So AI is simply trying to be a kind of indexical system which will more accurately guide us to something.
There is, that said, A caveat, you know, because I don't want to be too boosterish.
One of the really important things, you mentioned suffering.
One of the reasons why therapy and psychoanalysis even doesn't go on all the time, apart from the practical things, is there's an idea that we need to pause in order to learn.
That actually not thinking, in inverted commas, has a privileged role to play in getting certain ideas to gestate.
That we need fallow periods in which things are planted in us because of this thing which we call resistance.
So certain thoughts are so overwhelming that they need time for our defenses to lower and for us to accept them.
I mean, at a most basic level, it's why many of our best thoughts come to us in the shower or in the bath or on a walk, because we're not trying to think.
If you say to somebody, think now, you may not get the best thinking out of them because good thinking is associated with anxiety, and it's the anxiety of upturning.
so-called normality, which is where the repository of cliches and bad ideas and dead ends resides.
So in order to try something new, there is often a charge of anxiety, which is why we don't want to think.
And part of the problem of technology is that it gives us constant opportunities not to think.
I mean, in the most drastic sense, we can look at things that are completely irrelevant to us.
But even if we're using AI to roughly think in the area where we need to think, that thinking from outside can prevent the gestation of those inner thoughts.
So we do need to pause, we do need to sleep, and there is a serious danger that we won't give time to the really most important realizations.
Can we apply that to you and ask you how you are using AI in your life and where?
have you found it to be useful to you or pushed you?
And where have you found it to be failing?
Yeah, I mean, someone said to me, you know, it's got jagged intelligence.
In other words, at points, it is extremely good.
At other points, it's surprisingly bad.
And I think we're all still working out those surprising moments when, you know, it's much better than we expected, or it's much worse than we expected.
I have used it a lot in relationships.
I got divorced.
I went dating.
I got a new relationship.
All of these moments, AI has been...
been there.
And it's often had surprisingly good things to say.
I mean, I'll admit, as I say, I've done a psychotherapy training, I'm a practicing psychotherapist, I will put, say, hand on heart that at points, it beats me and my colleagues in terms of what it does.
The accuracy, the perspicacity, the capacity to connect up dynamics is at points, not always, but at points, extremely impressive.
And I think, you know, we need to recognize that.
And I think practicing therapists need to learn from AI.
and not to see it as their enemy, just as practicing writers do.
It's all about finding how to work with the technology and get the best out of oneself and the best out of the technology.
But it's certainly not useless.
I think I do see it as a perspective-giving tool, and it's particularly good at having a memory.
It's particularly good.
A lot of what psychotherapy really is is making connections.
You did something today.
How does that relate to something in your childhood, et cetera?
Pattern recognition is the forte of this technology.
And it's very close to the essence of what psychotherapy is.
But just staying on, so I want to talk about psychotherapy a bit more.
But have you found in your use of it that you have made better decisions or the result of your relationships has been better than, I mean, it's hard to compare what it would have been like without.
But in terms of your observations, has it helped?
Do you think it has made a positive impact on how you relate to the people there?
Yes.
I mean, look, as I say, it tends to deliver...
very cogent, very reasonable, very reassuring feces at moments of panic and at moments where perspective is lost.
So I don't see the problem there.
And I do see it as perspective giving.
It does depend on the quality of the prompts, of course.
Part of the problem is that we don't understand ourselves well enough to use it properly.
So we're not able to tell it the kinds of things that it needs to hear in order to make the proper...
connections.
I mean, when Freud was developing psychoanalysis, he came up with this very ingenious idea known as free association, which sounds very peculiar.
It still remains very provocative.
The idea is that you shouldn't sit like we are sitting, you know, face to face.
One of you should be lying down with your eyes possibly closed in order to do this very odd thing, which he called the fundamental rule, which is to say to the analyst everything.
truly everything that is passing through your mind at the moment that it does so.
And the idea is that there's constantly a battle in us between, as it were, the raw data that's in our minds and a normalizing, kind of sanitizing impulse to clear away from that raw...
stream of consciousness, those bits of our thinking that are troubling, that are contrary to our view of ourselves.
We very often have, all of us have a...
picture of who we would like to be.
And the reality of all of us is far more troubling.
Our sexuality, this was Freud's great insight, our sexuality is, on the whole, extremely provocative to most of our normal understanding of what a decent human being should be.
That if you actually tap into what people fantasize about...
think about sexually, you will get results that are really quite shocking.
I mean, we would all be shocked if we, you know, if we all put a microphone in everybody's head.
I mean, this is the paradox in this sense, the comedy of it.
If you put a microphone in anybody's head and they would sound extremely abnormal.
However, this is true of everybody.
So our sense of the normal is not normal enough.
It's not truly comprehensive.
And Freud's idea is that because we're constantly attempting to be normal, we're missing out on very rich and interesting data on which our actual flourishing depends.
If you're trying to do something creative...
you really need to get in touch with what you actually think, what you actually want.
If you're trying to decide on what your career should be, on how you would like to lead your life, very often these impulses and these desires are not obvious to us because they are challenging our settled view of ourselves.
And this is where psychoanalysis has traditionally made use of this couch and stream of consciousness to basically say, tell me everything without censoring yourself.
And on that basis, I, the therapist, will be able to pick up on things on which your true flourishing depends.
You know, the psychologist Donald Winnicott made this distinction between a true self and a false self.
He said that all of us have this kind of...
Outward shell of the full self where we say that everything's fine and we say we're flourishing, etc.
And then we have a true self the locus of our our true Feelings passions conflicts etc.
And the true self is is very often so buried beneath a kind of dutiful Sense of what others expect from us that we fall ill from that kind of tension and and the proverbial breakdown What we call a breakdown in psychotherapy is really a breakthrough Waiting to happen.
It's it's an attempt by some true element of us which hasn't found space during daylight hours to make its way Through to us now You can see that with AI this presents really interesting a really interesting conundrum, because if you use the AI to buttress your false self, you're going to get some very false results.
If you dare to use AI to reveal your true self, well, you may have privacy concerns.
You know, the raw data out of anyone's mind may, you know, get the FBI on us.
Who knows?
So, you know, it really does raise interesting questions of our sense of...
What is a dangerous human being?
What is a dangerous thought?
Because many of the thoughts that we have on which our sanity depends sound crazy.
So we need to find room for the crazy thoughts in order to be sane.
And we can do that in therapy.
Therapists know about that.
Therapy tries to create an environment in which that can happen.
Can AI therapy do that?
We're not there yet, but potentially it can do that.
Potentially it could do that even better, but so many issues still to address.
So you said before that human psychotherapists need to learn from.
AI, right?
Well, how do you think the training of therapists, you know, will be changed by AI or kind of how can AI augment?
Yeah.
Look, I think that what we'll see is what we're going to see in lots of areas, which is the weakest will be winnowed out and the strongest will get stronger.
And in the same way that, you know, people still pay for a handcrafted piece of clothing, even though a factory made piece of clothing can be had much cheaper, that there is some kind of benefit to the handcrafted thing.
Most of the time we'll be accepting the factory made, the automated version, and that will be really very good.
But there will still be a...
for a living and breathing psychotherapist.
But let's remember that most people cannot afford psychotherapy.
Psychotherapy is extremely expensive.
Mostly, you know, it's from £40 upwards.
In London, you could be paying £150 for 50 minutes.
Extremely expensive.
And, you know, simply out of reach of many, many people who need it.
So there is a real possibility of democratising something that has been out of reach.
It doesn't have to replace ordinary or real or in-person psychotherapy, but it can enhance it and work alongside it.
Okay, so what would be your sort of ideal view in five years' time of how we were making best use of this technology as individuals to be happy, to make the most out of our relationships and to learn to love better and be better at them?
Look, we are extremely confused creatures.
We're mostly living in the dark.
We mostly go to our deathbeds, not fully understanding who we are and who we've been.
We are a mystery.
to ourselves.
The philosophical adventure begins in ancient Greece with the command, know yourself.
And the idea that knowing yourself is both an enormous challenge and is also the gateway to flourishing.
I think AI is part of that story.
It doesn't have to be the enemy of that story.
I mean, you know, we'll probably find a way to mess it up.
Humans generally do.
But I'm trying to be slightly provocative from, you know, I come from a humanities background, which tends to be extremely caustic towards all technological discoveries.
I'm trying to almost deliberately look at the possibilities.
I'm not saying those possibilities will come true, but I want to highlight what those possibilities are.
AI, as part of that very long history of human beings, trying to understand themselves against all the obstacles that there have always been, fear, anxiety, a desire to fit in, a desire to be normal, when actually there's nothing to be afraid of, our vision of normality is false, etc.
And I think AI could be part of that story.
So use it as a mirror and reflect us back to ourselves.
That's very beautifully put.
Thank you so much.
Thank you.
Thank you.
That was me, Madhumita Merjia, speaking to Alain de Botton at the FT Weekend Festival.
Thanks to Alain.
And if you want more about the impact of artificial intelligence on our emotional lives, check out Artificial Intimacy, the six-part Tectonic series hosted by my colleague, Christina Criddle, earlier this year.
Scroll back on the Tectonic feed to find it.
The executive producer is Topher Forges.
Sound design by Breen Turner and Sam Giovinko.
The FT's head of audio is Flo Phillips.