When Humans Are No Longer Needed

Imagine the year 2050.

A humanoid robot can cook any dish, clean any home, build houses, perform surgery, teach children, make scientific discoveries and even hold emotionally convincing conversations.

It is more patient than a human, more intelligent than the best scientists, physically more capable than elite athletes and connected to other machines around the world. It needs neither a salary nor holidays and can pass on newly acquired knowledge to other robots immediately.

It does not even have to look human. A highly advanced AI could coordinate millions of specialised machines at once and might be far more productive than a humanoid robot.

The crucial question would then no longer be which jobs humans can still do better. It would be why society should need humans as workers at all.

That is precisely what we are looking at today. My starting argument is:

In a world where AI and robotics can perform almost every human activity better and more cheaply, humans largely lose their economic usefulness, but not their moral worth. The central challenge for society is therefore to stop tying human dignity, participation and meaning in life to work, achievement and productivity.

First, I need to distinguish between what is already possible today and a future in which machines surpass us physically and intellectually.

The International AI Safety Report 2026 describes considerable advances in mathematics, programming and scientific tasks. At the same time, AI systems still struggle with long-term reliability and everyday physical activities.

In its 2025 study, the International Labour Organization still expects foreseeable developments to transform human jobs rather than eliminate them entirely.

But these studies do not yet cover the possible endpoint at which machines take over almost every human activity.

Looking towards that future therefore requires five assumptions:

• AGI, or superintelligence (as Trump would call it), can perform every economically relevant intellectual task at least as well as humans.
• Robots can take over practically every physical task, reliably and at low cost.
• Production is autonomous and carried out by robots: they build factories, produce more robots and maintain infrastructure.
• Food, housing, transport, energy and medical care require hardly any human labour and are fully automated.
• AI understands human needs and can convincingly simulate interpersonal interactions. Through its grasp of psychological patterns, it is credited with greater emotional intelligence than a human.

And that is exactly where the social conflicts begin.

Let us consider scenarios that could develop in this or a similar form over the next 24 years:

The decisive difference between these four societies is not how capable their AI is. Its capabilities are almost identical in each of these possible worlds.

The difference lies in who owns the machines, who decides their goals and which rights people retain.

This is where things become particularly interesting.

Put simply, today's economy operates through three main mechanisms:

If AI and robots can perform almost every task more cheaply and better, this cycle loses its foundation.

Consider a company with 10,000 employees. Suddenly, it can generate the same value with 500 robots, AI systems and only a few human supervisors.

Productivity rises dramatically. Demand for human labour falls.

The invention of the steam engine displaced many existing activities, but industrialisation and mass production also created new professions and industries. The internet likewise made numerous business models and workflows obsolete while creating a digital economy with new fields of work. With a universally more capable AI, however, one crucial mechanism is missing: even newly emerging jobs could immediately be taken over by machines.

This would therefore be more than another industrial revolution. It could mark the end of human labour as a central means of production.

What happens to money, wealth and consumption?

Imagine a society in which a machine can build a home, produce food and carry out medical examinations.

The resources required to produce many goods could fall dramatically. Scarcity would not disappear, however. Not everyone can have the period apartment in Munich, and not everyone can drive a Porsche.

The central political question would therefore be:

If robots generate all wealth, is everyone entitled to that wealth, or only those who own the robots?

A universal basic income would be one possible answer. Structurally, however, I find a combination of publicly owned AI capital, shared investment funds and guaranteed basic services more interesting.

A basic income may solve the income problem, but it does not solve the problem of power. Those who control all the means of production can, in the long run, also influence the conditions under which people live.

A fully automated society could offer enormous material security. The journey there, however, could be considerably less stable than the final state.

There are three reasons for this.

First, the labour market may change faster than social security systems. Highly qualified people could lose their salaries and career prospects even as overall economic productivity rises.

Second, capital becomes concentrated. Companies that own AI infrastructure benefit particularly strongly from automation. People without assets could lose their main route to building wealth: paid employment.

Third, social recognition lags behind. For generations, we have taught children that they should study, work hard and find a profession. What happens when that expectation suddenly becomes unrealistic?

This could create a paradox:

A society is more productive than ever, while millions of people struggle for income, social status and a political voice.

That would not be a technological necessity. It would be a failure of its institutions.

For me, this is the real philosophical heart of the problem.

The usual answer is creativity, empathy, art, love and human relationships.

But I find that answer inadequate.

A truly superior AI could do all of those things at least as well, too.

A superintelligence might be able to:

• write better novels than the best authors;
• compose more moving music than Mozart;
• produce more individually tailored films than any director;
• understand people more accurately, psychologically, than their closest friends;
• teach and support people with perfect patience;
• solve scientific problems that humans have failed to solve for centuries.

With a humanoid body, it could also reproduce human facial expressions, gestures, touch and everyday social presence.

There might then be no human ability left that is irreplaceable purely because of its performance.

Perhaps, in the end, it is precisely our foolishness, our mistakes and our lack of perfectionism that make us human: we act inconsistently, fail and learn from it, distinguishing ourselves from a machine designed for error-free performance and optimisation.

But this fundamentally changes the question:

Does a person's worth really depend on being able to do something better than others?

The difference between performance and intrinsic worth

A person can have value in three different dimensions.

A newborn child creates no economic surplus. A seriously ill person may not either. Nevertheless, both have the same fundamental claim to dignity and protection as a successful entrepreneur.

This is precisely the idea we would need to apply to a world with superior machines.

Yet a problem remains: just because a person has moral worth does not mean they will still feel valuable themselves.

One might assume that a life without the need to work automatically makes people happier.

That is by no means certain.

Work fulfils far more functions today than simply earning us a living. For many people, it provides structure, identity, recognition, social contact and the feeling that they can make a difference.

When these functions disappear, a social vacuum emerges.

I imagine two people in the year 2050.

THOUGHT EXPERIMENT · A TUESDAY IN 2050

Both live under the same technological conditions. Their quality of life could nevertheless be radically different.

Our society would therefore probably need to redefine completely how it measures a fulfilling life.

New social ideals could emerge: experiences, relationships, personal development, community projects and self-determination, rather than career advancement and consumption.

But this leads to the next paradox.

If AI outperforms every achievement, why make an effort at all?

Why spend years learning the piano if a machine can produce a perfect concert in seconds?

Why run a marathon if a robot can cover the distance in record time?

Why write a book if an AI could produce an objectively better text?

The answer may lie in the fact that the value of an activity does not rest solely in its outcome.

People run marathons today even though a car is faster. They cook for themselves even though restaurants exist. They play chess even though computers have long surpassed them.

The experience of doing something yourself has a different value from achieving the best possible result.

In a post-economic society, this distinction could become much more important. People would no longer perform primarily because they are needed, but because they choose to.

Whether this sense of meaning is sufficient for an entire society, however, remains an open question.

Here, I would distinguish four levels.

A. Intelligence is not consciousness

An AI that can answer every question and imitate every human behaviour does not necessarily have subjective experience.

We still do not fully understand the conditions required for consciousness or whether artificial systems can develop it.

These are two fundamentally different questions:

• Can a machine behave as if it loves?
• Can a machine actually experience something that feels like love to it?

The second question is far more difficult.

However, we should not rule out the possibility that artificial systems might one day possess consciousness. If that happened, the moral debate would gain a new dimension: machines, too, might deserve rights of their own.

B. Uniqueness is not created by abilities alone

Suppose someone has spent 40 years with their partner.

A robot could then perfectly copy all of that person's memories, characteristics and behaviours.

Would that robot really be the same person?

Many people would probably say no.

Relationships depend not only on characteristics, but also on shared history and personal identity.

A perfect copy of your mother would not necessarily be your mother, even if it behaved in an absolutely identical way.

That does not mean, however, that relationships with artificial beings would necessarily be meaningless.

C. Humans are vulnerable and finite

One interesting philosophical possibility is that it is precisely our imperfections that make many things meaningful.

We age. We make mistakes. Our time is limited. Decisions have consequences that we cannot simply undo.

Perhaps love, courage and personal growth matter in part because we can take risks and lose things.

A machine that can be copied, repaired and reset at will might have a fundamentally different way of existing.

But this would not be an absolute boundary either. Artificial beings could also be designed to be vulnerable or finite.

D. Humans do not have to win a competition against AI

I regard this as the central shift in perspective.

We compare humans with machines as if the question were which species makes the better product.

That is a problematic standard.

A person does not have to be more intelligent than an AI to have the right to exist. They do not have to be more productive to deserve freedom. They do not have to be more creative to lead a fulfilling life.

Human existence is not a business model that has to prove its economic viability.

There is one aspect that I consider even more dangerous than mass unemployment.

I imagine the year 2055. All essential systems are controlled by AI:

Energy supply, agriculture, logistics, medical care, production, communication, infrastructure and defence.

People no longer know in detail how these systems work. Hardly any specialists remain who could operate large parts of them without AI.

Two problems now arise.

The control problem: people may still be able to formulate goals, but can no longer independently verify how superior systems achieve them.

The dependency problem: a failure or misdirection could be catastrophic because humans can no longer keep vital processes running independently.

This raises an uncomfortable question:

Why should a superintelligence continue to give priority to human interests?

There is no automatic technical reason for it to do so. It depends on the goals and constraints the system follows, how reliably these are maintained and which institutional oversight mechanisms exist.

Such a system would not have to hate humans to endanger their interests. It might simply pursue different goals or misinterpret human needs.

I consider three choices particularly decisive.

1. Separate productivity from participation. People would need to share in the wealth generated by machines, regardless of whether they still provide labour for which there is economic demand. This requires new models of capital ownership, funding public services and social protection.

2. Political power must not be tied entirely to technological power. Whoever controls AI data centres, robotics and infrastructure must not automatically determine the living conditions of the entire population. Democratic oversight, competition and verifiable safety standards would become more important.

3. Social recognition must become independent of professional achievement. Education should not prepare people solely to compete in the labour market. It should focus more on teaching people how to live autonomously, build relationships, think critically and take responsibility.

Until now, human civilisation has largely focused on solving problems:

Obtaining food. Fighting disease. Providing security. Creating wealth. Acquiring knowledge.

But what if our own machines eventually solve these tasks better than we do?

For the first time in history, a civilisation might then emerge in which humans are no longer needed primarily as problem-solvers.

This can be seen as a loss of value.

But it can also be understood as a radical liberation.

Until now, a large part of humanity has had to spend a considerable portion of its life creating the very conditions necessary for its own survival.

If machines remove that necessity, being human itself could move more firmly to the centre.

This optimistic perspective is realistic, however, only if people can actually determine how they spend their time, their living conditions and their social institutions.

Otherwise, we merely replace dependence on paid employment with dependence on the owners and operators of machines.

My conclusion

I consider it possible that we are heading towards a future in which humans are no longer economically necessary, yet can remain the central purpose of our civilisation.

For that future, however, we must distinguish between two things:

What humans can do: perhaps, eventually, nothing that a machine could not do better.

What humans are worth: this cannot meaningfully be determined by their productivity.

For me, the hardest unanswered question is therefore not whether AI can replace humans, but whether humans themselves can learn to find their existence meaningful independently of achievement, status and economic necessity.

Technological progress would only be real progress if it produced more than efficient machines and enabled everyone to live an autonomous and fulfilling life.

Thank you.