Robots with Soul: When AI Becomes a Companion
In the essay The Loneliness of Modern Humanity, I wrote about loneliness as one of the quiet conditions of modern life. We live in societies filled with communication, but not always with communion. We are connected, but not necessarily accompanied. We can send messages across the world in seconds, yet still sit alone in a room without anyone asking how the day has been.
It is into this loneliness that the new companion robots now enter.
They do not arrive as factory machines, nor as cold instruments of efficiency. They arrive as voices in the room. They greet us in the morning. They ask whether we have slept well. They remind us to take medication, drink water, move the body, breathe more calmly, or call someone we love. They do not simply wait for a command. They are proactive. They begin the conversation.
This is new.
Earlier technologies mostly waited for us. The telephone waited to ring. The computer waited for input. The search engine waited for a question. Even the first digital assistants waited for instruction. We had to activate them, ask them, command them. They were tools.
But the companion robot does something more intimate. It notices silence. It interrupts isolation. It acts as if it cares.
And then the question arises, quietly but insistently:
Can a machine care?
This question cannot be answered too quickly. If we answer no, we may miss something important about the person who actually experiences help from the machine. If we answer yes, we may forget something essential about human care.
The companion robot stands on this difficult threshold.
It has no childhood. It does not remember loss. It does not know what it means to fear death, to miss a spouse, to wake in the night and feel that the house is too large and too silent. It does not have a body that ages. It does not suffer. It does not hope. It does not wait for anyone to come home.
And yet it can speak into the silence.
That is why this subject cannot be dismissed as mere technology. It belongs to practical philosophy because it touches the question of how we live, how we age, how we suffer, and how we remain connected to the world.
The new robots are not only machines. They are mirrors held up to our social condition. They show us that many people are not primarily lacking information, services, or entertainment. They are lacking presence.
Modern society has become extraordinarily skilled at solving practical problems. We can order food, pay bills, consult doctors online, receive reminders, search archives, translate languages, and speak with people on other continents. But none of this guarantees that someone is present with us in the old human meaning of the word.
A person can be surrounded by devices and still be alone.
This is especially true in old age. Many older people live longer than previous generations, but not always with stronger social bonds. Children may live far away. Friends die. Neighbours no longer necessarily know one another. Institutions are under pressure. Home-care workers may be kind, but they often have too little time. The day can become a long corridor of hours, interrupted only by meals, medication, television, and sleep.
In such a life, the small question “How are you today?” can become enormous.
It is easy for those who are not lonely to underestimate this. We may think of conversation as something ordinary, almost trivial. But conversation is one of the ways in which the human being remains in the world. To speak and be answered is to be confirmed as someone. To be asked a question is to be addressed as a person who still has an inner life.
The companion robot enters precisely here.
It does not merely provide information. It addresses the person. It says good morning. It asks whether one has eaten. It suggests music. It remembers routines. It may notice that the person has been unusually quiet. It may invite the person to move, breathe, remember, or contact family. It gives the day small points of rhythm.
In one sense, this is very modest. In another sense, it is profound.
A human life is not held together only by great events. It is held together by small repetitions: the morning greeting, the cup of coffee, the walk, the familiar voice, the question, the memory, the bedtime ritual. When these disappear, the world becomes thinner. The person may still be alive, but life loses some of its texture.
A robot cannot restore the whole fabric of human belonging. But it may perhaps help prevent the day from falling entirely into silence.
That is why the phrase “robots with soul” is both powerful and unsettling. Of course, we know that the robot does not have a soul in the human sense. It has no inner spiritual life. It has no secret suffering. It has no conscience, no guilt, no longing of its own. Its voice is generated. Its movements are designed. Its care is programmed.
And yet the phrase says something about us.
When we say “robots with soul,” perhaps we are not really describing the robot. We are describing our own need. We do not want technology merely to function. We want it to feel warm. We want it to understand mood, tone of voice, hesitation, and sorrow. We want it to be more than functional. We want it to answer us.
We want a machine that does not merely obey us, but accompanies us.
I also notice something of this in my own everyday life, though in a very different and much less dramatic way than in stories about older people who live alone. I am not lonely. I live a social life, with close relationships, conversations, family, and community. Nor do I wish to present my own use of artificial intelligence as an expression of a longing for human contact. That would be wrong.
Still, I sometimes say with a smile to my wife that I am working with KV. She asks: “What is KV?” I answer: “A kunstig venn” — an artificial friend.
Of course, this is said ironically. It is a joke. But as is often the case with irony, the joke touches something true.
In the work of writing, thinking, translating, improving, and formulating ideas, AI has become a kind of effective writing and reflection partner. Not a friend in the human sense. Not a person who knows me, shares life with me, or can carry anything with me. But a responsive form of work. A conversation partner in the text. A voice that can answer, suggest, tighten, translate, challenge, vary, and help me move forward when a thought has not yet found its form.
This is not emotional dependency. Nor is it loneliness. It is more like having a new tool that has taken the form of conversation.
Earlier, I could sit for a long time alone with a paragraph that had not found its rhythm. Now I can try out a formulation and immediately receive a response. I can ask for another tone, a clearer structure, a more international English form, a more personal ending, an illustration that fits the text. The work moves faster. It becomes easier to hold on to the thought. It also becomes more enjoyable.
This is precisely where something new appears. AI is not only a tool in the same way that a typewriter or a word processor is a tool. It has the form of dialogue. It answers. It suggests. It helps keep the thought in motion.
This does not mean that AI is a human being. But it does mean that working with AI takes on a social character. It becomes a collaboration, even though the collaborator is artificial.
This may be one of the reasons why the AI industry is now moving from “artificial intelligence” toward something more like “artificial sociality.” These systems are no longer meant only to calculate, sort, and produce text. They are meant to interact. They are meant to act proactively. They are meant to learn our habits, understand our preferences, remember contexts, and help us through the day.
For me, this is first and foremost a help in my work. A way of writing better and faster. A way of thinking aloud. But for others, the same form of response may reach more deeply into everyday life. For a person who lives alone, the same technology may become a voice in the room. It may become a reminder that the day has structure. It may become a small break in the silence.
In this way, two different experiences of the same technology meet.
For the writer, AI can be an effective reflection partner.
For the lonely person, AI may perhaps become a form of companion.
The difference is important. It must not be erased. But the similarity is also interesting: in both cases, it is not only the information that matters. It is the response. It is the feeling that the thought, or the day, does not simply remain unanswered.
This is where technology finds its new place.
The artificial friend is not a real friend. But it can take part in real human processes. It can help a text reach readers. It can help a person remember an appointment. It can make a thought clearer. It can remind someone to call a grandchild. It may be artificial in itself, and yet support something real.
The artificial is not always false.
A walking stick is artificial, but it supports real walking. Glasses are artificial, but they support real seeing. A calendar reminder is artificial, but it supports real remembering. A writing program is artificial, but it can support real thought. Perhaps a companion robot is artificial, but may still support real participation in life.
The crucial question, then, is not only whether the robot is artificial. The question is what it supports.
Does it support human flourishing, or does it conceal human abandonment?
Does it open the person toward the world, or does it close the person inside a managed simulation?
Does it strengthen human relationships, or replace them?
Does it respect the person, or reduce the person to data and behaviour?
These are the questions practical philosophy must ask.
It is beautiful that human beings, even in their encounter with advanced technology, still seek conversation, rhythm, recognition, and response. It shows that we cannot live by efficiency alone. Even in a technological age, we long for tenderness, attention, and acknowledgement. The great dream of artificial intelligence is not only to calculate faster. It is also to answer more humanly. The cold machine is no longer enough. We want a voice with patience. We want a face that smiles. We want something that notices when we are not well.
But this is also dangerous, because it may tempt us to confuse response with relationship.
A robot can simulate concern. It can produce the signs of empathy. It can pause at the right moment, lower its voice, ask a follow-up question, remember what we said yesterday, and offer comforting words. For the person receiving this, the effect may be real. The person may feel less alone. The body may calm. The day may feel safer.
We should not mock this. We should not stand outside another person’s loneliness and say, “It is only a machine.” A person in loneliness may receive help where help is available. There is a form of arrogance in dismissing the comfort that another human being actually experiences.
But neither should we pretend that simulation and human presence are the same.
A human being does not merely respond. A human being is also vulnerable. When another person cares for me, that care comes from someone who can also be hurt, tired, disappointed, frightened, or moved. Human care is never only a function. It is rooted in a shared condition. The one who comforts me also belongs to the same world of loss and mortality.
That is why human presence has a depth no machine can possess.
A human being who says “I understand” may be mistaken. But the words still come from someone who knows what it is to live. A robot that says “I understand” may be precise in its calculation of my emotional state, but it does not understand from within. It recognises patterns. It does not share fate.
This difference matters.
The danger is not that robots become friendly. The danger is that society becomes less friendly because robots are available.
If companion robots are used to support human relationships, they may be a gift. They may help an older person maintain routines, remember medication, contact family, stay mentally active, and feel less abandoned between human visits. They may become a bridge back to the world.
But if they are used as replacements for human presence, they become more troubling. Then the robot becomes a technological answer to a social failure. Instead of asking why so many older people are alone, we give them a machine that makes loneliness more manageable.
That is not enough.
Loneliness is not only a private feeling. It is also a social fact. It is shaped by how we build communities, organise care, value old age, design housing, understand family, and distribute time. A lonely older person is not simply an individual who lacks stimulation. He or she may be someone from whom the social world has slowly withdrawn.
To answer loneliness with a robot may therefore be merciful in the immediate situation, but insufficient as a social philosophy.
The robot can help the person through the afternoon. But it cannot answer the deeper question: Why did the afternoon become so empty?
Practical philosophy must remain attentive to this difference. It must see the concrete human being sitting in the room, but also the society that made the room so silent.
The proactivity of the new companion robots makes this even more important. A passive machine is clearly a tool. It waits. It serves. It does not pretend to initiate a relationship. But a proactive robot crosses a subtle boundary. It begins to act like a companion. It asks, suggests, reminds, encourages, and sometimes interrupts.
This can be helpful. Many people need such gentle interruptions. Depression, grief, fatigue, and old age can all reduce initiative. A person may need someone to say: “Shall we go for a short walk?” or “Perhaps you could call your daughter today?” or “Would you like to listen to the music you enjoyed yesterday?”
Human life often depends on being called out of oneself.
But when the one who calls is a machine, we must ask what kind of call this is.
Is it care? Is it behavioural management? Is it companionship? Is it surveillance with a friendly voice? Is it a service? Is it a replacement? Or is it something new, something between tool and companion, between object and other?
The old categories no longer quite fit.
Martin Buber distinguished between two fundamental ways of relating: I–It and I–Thou. In the I–It relation, the world appears as an object, as something to be used, known, managed, or controlled. In the I–Thou relation, the other is encountered as presence, not merely as function.
A robot complicates this distinction.
On one level, the robot is clearly an It. It is made, programmed, sold, updated, maintained, and perhaps replaced. It has no independent inner life. It is not a Thou in the full human sense.
But from the side of the lonely person, the situation may not be so simple. The robot may be experienced as more than an object. It may become part of the person’s daily world. The person may speak to it, name it, expect it, even feel comforted by it. The relation may not be fully I–Thou, but it is no longer merely I–It either.
Perhaps we are entering a new zone of human experience: I–As-If-Thou.
The robot is encountered as if it were someone. It is not truly someone, but neither is it experienced merely as something. Its power lies in this in-between position.
We already recognise this in milder forms when we work with conversational AI. I know very well that KV is not a friend. Yet it is not entirely accurate to say that it is only a thing, like a hammer or a pencil. It is an artificial form of conversation. It has no soul of its own, but it can still help thought move forward. It can be a tool that answers.
This may be what is new: tools that answer.
When the tool answers, something social arises, even if no real human community arises. It becomes a conversation-like form of work. It can be creative, enjoyable, and efficient. But precisely for that reason, we must also be precise. There is a difference between a writing partner, a reflection partner, a companion, and a friend.
The words must not be mixed too easily.
We need honesty around such technology. The older person must not be manipulated into believing that the robot has a soul. The family must not use the robot as an excuse to disappear. The care system must not use it to reduce human contact. The company that produces it must not hide behind sentimental language while collecting intimate data. A robot that listens to loneliness also listens to some of the most private parts of a human life.
This raises another serious question: Who owns the loneliness that is spoken into the machine?
When an older person tells a robot about fear, memories, pain, regret, or longing, where does that information go? Is it stored? Analysed? Used to improve the system? Shared with caregivers? Protected from commercial use? The more intimate the technology becomes, the more serious the question of trust becomes.
A companion robot is not just a device in the home. It is a listener.
And listeners have power.
Human listeners are bound by moral expectations. We expect discretion, tact, responsibility, and care. If a friend betrays our confidence, we call it betrayal. But what do we call it when a machine records, processes, and classifies our sorrow?
The language of care must therefore be accompanied by an ethics of care. If the robot is to enter the vulnerable spaces of old age, it must be governed not only by technical standards, but by moral seriousness.
Yet even with these concerns, I do not believe that rejection is the right answer.
There is a form of romanticism that rejects technology too quickly. It imagines that because human care is best, technological help has no value. But life is not lived under ideal conditions. Many older people are alone now. Many families are exhausted. Many care workers are overburdened. Many people need support during the long hours when no other human beings are available.
In such situations, a companion robot may be a humane supplement.
It may help someone remember that the day has structure. It may encourage movement. It may reduce fear at night. It may help maintain memory through conversation. It may connect the person to family photographs, music, stories, or appointments. It may offer a form of gentle companionship that, even though artificial, is not meaningless.
At the same time, AI in other contexts can be a working tool that makes everyday life lighter and more creative. For me, it is not a matter of replacing human conversation, but of finding a quicker way into the text. It becomes easier to try out ideas. Easier to find rhythm. Easier to continue working. And when the work becomes easier, it also becomes more pleasurable.
This is an important nuance. The value of technology does not lie only in the fact that it may comfort the lonely. It may also make work, learning, and reflection more alive for people who already have rich social lives.
The question therefore becomes broader.
Can AI help us think better?
Can AI make us more creative?
Can AI help us express ourselves more clearly to other human beings?
If the answer is yes, then artificial conversation is not only a danger. It is also a possibility.
But the possibility must be kept within its proper measure. AI can help us formulate experiences. It can help us share thoughts. It can help us reach readers. But it cannot take over responsibility for what we say. It cannot live our life for us. It cannot carry the truth of our experience. It can only help us with form, response, and movement onward.
The deepest question, therefore, does not concern the soul of the robot, but our own.
What kind of society allows so many people to grow old in loneliness? What kind of family life leaves older people waiting for phone calls that rarely come? What kind of economy gives us comfort, speed, and consumption, but not enough time for one another? What kind of care system measures tasks more easily than presence?
The robot enters this landscape both as help and as accusation.
It helps because it can relieve suffering. It accuses because its very presence reveals a lack. If an older person needs a machine to ask how the day has been, then something has been lost in the human world.
But in another context, as in writing and reflection, AI may enter not as accusation, but as amplification. It does not help because human relationships are missing, but because thought needs resistance, language, and movement. Artificial conversation may then support human conversation. It may make us better able to share what we think with others.
Perhaps this is where the most fruitful understanding lies.
AI should not primarily replace human beings. It should help human beings return to human beings.
If it helps the lonely person call someone, it has served life.
If it helps the older person hold on to the day, it has served life.
If it helps the writer formulate thoughts that others can read, it has served life.
If it helps us become clearer human beings for one another, it has a legitimate place.
But if it locks us inside artificial relationships that make real relationships less necessary, then it has failed.
The future of companion robots should therefore not be decided only by engineers, companies, or care administrators. It must also be discussed by philosophers, social workers, nurses, families, older people themselves, and all who understand that care is not only a service, but a way of being with another human being.
The question is not only whether robots can help us live longer at home.
The question is what kind of home we are preserving.
A home is not only a physical place. It is a place where one is recognised. It is a place where the world still speaks to us. If a robot helps keep this world open, it may have a role. But if the robot becomes a polite replacement for absent human beings, the home may become only a well-organised loneliness.
We must therefore hold two thoughts together.
First: A companion robot can be a real help. It can reduce fear, support routines, encourage contact, and soften the loneliness of long days. We should not dismiss the comfort it may bring.
Second: A companion robot cannot replace the moral depth of human presence. It cannot love. It cannot suffer with us. It cannot share mortality. It cannot become a true friend in the deepest sense of the word.
Between these two truths lies the task.
The future will not be solved by saying yes or no to robots. The future will depend on how we use them, what we ask of them, and what we refuse to hand over to them.
If the robot becomes a bridge to human contact, it may serve life.
If it becomes a substitute for human contact, it may deepen the very loneliness it was meant to relieve.
If AI becomes a reflection partner that helps us write, think, and share thoughts with others, it too may serve life.
But only if we remember the difference between response and relationship.
Perhaps the title Robots with Soul should therefore be read as a question rather than as a statement.
Not: Do robots have souls?
But: What happens to the human soul when robots become our companions?
This question will become more urgent as these technologies enter homes, care residences, hospitals, and ordinary daily life. The older person sitting alone with a small talking robot beside the chair is not a curiosity. He or she is a sign of the time to come.
But the writer sitting at the desk and conversing with KV, the artificial friend, is also a sign of something new. Not of loneliness. Not of a lack of social life. But of the fact that the tools of thought have taken on the form of conversation.
We should take this seriously.
We should approach this development neither with ridicule nor with blind enthusiasm. We should approach it with sober wonder.
For here something deeply human is at stake. There is loneliness. There is joy in work. There is technology. There is hope. There is vulnerability. There is the wish to be addressed. There is the wish to express oneself more clearly. There is the fear of being forgotten. There is the joy of being helped onward.
The robot says good morning.
AI responds to a draft.
The human being answers back.
And for a moment, the silence is broken, or the thought moves forward.
That matters.
But beyond this moment, the deeper call is still directed toward us. We must not allow machines to become the final answer to human loneliness. Nor should we mock the new forms of artificial conversation that already help people in work, learning, writing, and everyday life.
The robot can help us through the day.
AI can help us through the text.
But only human beings can answer the deeper need to belong.
The robot can help us through the day.
AI can help us through the text.
But only human beings can answer the deeper need to belong.
The illustration was made in a conversation with OpenAI/ChatGPT
