Work’s Worth
Is There a Human Right to Human Work?
Artificial intelligence is transforming the debate about the future of work.
Much of that debate is economic. Which occupations will be automated? How many jobs will disappear? Which new jobs will be created? What happens to wages and productivity?
These are important questions. But there is another question that receives considerably less attention:
What happens to the human right to work when machines can increasingly perform the work that human beings once did?
This is not simply a question about unemployment.
Work provides income, but for many people it also provides something more: independence, purpose, social participation, the development and exercise of skills, and the experience of contributing to a community.
International human rights law already recognises some of these interests. The challenge posed by artificial intelligence may therefore require us to look once more at what the right to work actually protects and whether our existing understanding of that right is adequate for a world in which human labour is increasingly technologically replaceable.
There is already a human right to work
The right to work is not a new idea.
Article 23 of the Universal Declaration of Human Rights recognises the right to work, to free choice of employment, to just and favourable conditions of work and to protection against unemployment.
Article 6 of the International Covenant on Economic, Social and Cultural Rights (ICESCR) goes further. It recognises:
“the right of everyone to the opportunity to gain his living by work which he freely chooses or accepts.”
The right is accompanied by obligations on States to take steps towards its realisation, including through policies, training and measures directed towards productive employment.
At European level, Article 1 of the European Social Charter requires States to pursue high and stable levels of employment, protect the worker's ability to earn a living in an occupation freely entered upon, maintain employment services and promote vocational guidance and training.
Within EU law, Article 15(1) of the Charter of Fundamental Rights of the European Union provides that:
“Everyone has the right to engage in work and to pursue a freely chosen or accepted occupation.”
None of these provisions means that every individual has an enforceable right to be provided with a particular job.
But neither is the right to work simply an aspiration that people should have an income.
The UN Committee on Economic, Social and Cultural Rights has described the right to work as an inherent part of human dignity. In its General Comment No. 18, the Committee explains that freely chosen or accepted work contributes not only to an individual's survival and that of their family, but also to their development and recognition within the community.
That distinction is enormously significant in the age of AI.
Why do human beings have a right to work?
If the purpose of work were exclusively to generate income, the technological displacement of human labour would pose a comparatively straightforward policy problem.
A sufficiently wealthy society could, at least theoretically, separate income from employment. Productivity gains generated by artificial intelligence could be redistributed through taxation, social protection, universal basic income or some other mechanism.
People might no longer need to work in order to live.
But that does not fully answer the human-rights question.
The existing right to work suggests that something other than income is at stake.
Work can provide a domain in which people exercise judgement, develop capabilities, acquire expertise, cooperate with others, undertake responsibility and contribute to society. Occupations can become important components of identity and social recognition.
This does not mean that all work is fulfilling. Much human labour is repetitive, dangerous, exploitative or poorly paid. Human rights law itself recognises this through the concept of decent work and through rights relating to working conditions, remuneration, equality and worker protection.
Nor should work become the sole measure of human worth. Caregiving, family life, education, community participation, artistic activity and countless other forms of human activity can provide meaning and opportunities for development outside paid employment.
The point is narrower.
If work is one of the principal social institutions through which people exercise agency and participate in collective life, then removing human beings from work raises questions that cannot necessarily be answered by replacing their lost wages.
AI changes the nature of the problem
Technological change has displaced human labour before.
Mechanisation transformed agriculture. Industrialisation replaced many forms of manual labour. Computers automated administrative and computational tasks.
It would therefore be a mistake to assume that every instance of technological displacement constitutes a human-rights problem.
Artificial intelligence nevertheless introduces distinctive questions because of the range of human activities it can potentially perform.
Generative AI systems can already produce text, images and software code; summarise and analyse information; assist with professional reasoning; perform administrative functions; and undertake components of occupations previously associated with highly educated human workers.
The International Labour Organization's 2025 assessment of generative AI and employment estimated that one in four workers globally is employed in an occupation with some degree of exposure to generative AI. Importantly, the ILO does not predict the wholesale disappearance of these jobs. Its assessment is that transformation is considerably more likely than complete replacement.
That distinction should shape the human-rights debate.
The immediate challenge may not be a world in which millions of occupations simply disappear.
It may instead be a world in which human beings remain employed while progressively fewer meaningful components of their work remain theirs to perform.
Job displacement is not the only issue. Task displacement matters too.
Imagine that an employee retains her job title and salary.
Before the introduction of AI, she researched problems, exercised professional judgement, drafted work, made decisions and developed expertise through repeated practice.
After AI is introduced, a system conducts the research, produces the first analysis, generates the draft and recommends the decision. Her principal function becomes checking and approving its output.
She has not lost her job.
But something about her work has changed.
That change may be beneficial. AI may remove repetitive tasks, increase productivity and give workers more time for activities requiring creativity, interpersonal engagement or sophisticated judgement.
But the opposite trajectory is also possible.
If AI systematically takes over the tasks through which people learn, reason, create, exercise judgement and develop professional competence, humans may retain employment while losing important opportunities for agency within employment.
The relevant question therefore becomes not simply:
Will AI replace workers?
It is also:
Which parts of human work are being transferred to machines, and what happens to the person when they are?
This is therefore also a question about the quality and structure of work, not merely its existence.
In a recent essay on the societal consequences of AI, Bill Gates proposed that governments consider designating certain occupations or tasks as “Human Reserved”. Certain areas would be deliberately preserved for human beings even where AI or robotics could technically perform the work.
Gates compares the idea to a nature reserve: we sometimes choose not to develop something simply because what would be lost is worth preserving. His examples range from caregiving to particular forms of human judgement and interaction.
The proposal raises a deeper question than whether automation should be slowed to protect employment. Are there forms of work whose value lies partly in the fact that a human being performs them? If so, the case for preserving human participation may rest not merely on economic protectionism, but on the value we attach to human agency, relationships, dignity and participation themselves.
There cannot be a general right against automation
There are strong reasons to resist an overly expansive interpretation of the right to work.
A human right to prevent machines from performing human tasks would be extremely difficult to justify.
Technological progress can eliminate dangerous and degrading labour. It can improve productivity, reduce costs, expand access to services and create entirely new forms of economic activity.
There is no obvious human-rights reason why a person should spend a working life performing a monotonous task simply because previous generations had to do so.
Nor can human rights plausibly guarantee the preservation of particular occupations indefinitely. Economies change. Consumer preferences change. Technologies change. The right to work has never meant that every existing job must survive those changes.
The question is therefore not whether people have a right to be protected from automation but rather whether the interests underlying the right to work impose limits or obligations concerning how technological transformation is managed.
That is a considerably more defensible proposition.
From a right to employment to a right to meaningful economic agency?
International human rights law already gives us clues.
The ICESCR protects work that is “freely chosen or accepted.”
The EU Charter protects the ability to pursue a “freely chosen or accepted occupation.”
And the UN Committee's interpretation of the right explicitly connects work with dignity, personal development and recognition within the community.
Taken together, these principles suggest that the protected interest cannot be reduced entirely to the receipt of wages.
They point towards a conception of the worker as an agent.
That does not establish a new legal right to “meaningful work”, still less a right to reserve particular tasks exclusively for humans.
But it raises a question that governments, employers and technology developers should begin taking seriously:
Should the governance of AI in the workplace consider the preservation and development of human agency as one of its objectives?
If the answer is yes, the implications extend beyond unemployment policy.
What might a human-rights approach require?
A human-rights approach to AI and work need not begin with prohibitions on automation.
It could begin with governance.
Before introducing systems that substantially restructure occupations, organisations could assess not only efficiency and financial return but their effects on workers - a worker impact assessment.
Which tasks are being automated?
Which tasks remain under meaningful human control?
Does the technology augment workers' capabilities or progressively substitute for them?
Will junior workers still have opportunities to acquire the skills that experienced professionals developed by performing tasks now delegated to AI?
Does the system increase human decision-making capacity, or reduce the worker to supervising automated outputs?
Are productivity gains improving working conditions, reducing unnecessary labour and creating opportunities for more valuable human activity or simply increasing the amount of output expected from each worker?
And where displacement does occur, are workers given realistic opportunities for retraining and transition?
These questions sit comfortably alongside existing human-rights obligations concerning work, equality, dignity and social protection.
They also suggest that the familiar distinction between augmentation and replacement deserves closer scrutiny.
The fact that a worker has not technically been replaced does not necessarily mean that technological transformation has preserved the human interests associated with work.
AI could make work more human
There is another possibility.
Artificial intelligence could strengthen rather than diminish the human dimension of work.
Many occupations contain tasks that are repetitive, administrative or cognitively draining but contribute relatively little to the aspects of work people find meaningful.
If AI performs those functions while leaving people with greater time for judgement, creativity, relationships, craftsmanship, problem-solving and care, technological transformation could advance some of the very values underlying the right to work.
The objective should therefore not be the preservation of human labour for its own sake.
It should be the preservation and expansion of human agency.
A society committed to human rights should not require people to perform unnecessary work simply because work has historically been necessary for survival.
But nor should it assume that once income is secured, nothing is lost when human beings cease to participate meaningfully in productive and professional life.
The question we need to ask now
The future of work is often presented as a competition between humans and machines.
That may be the wrong frame.
The more important question is what role technology should play in a society organised around human dignity and agency.
International human rights law already recognises a right to work. It connects that right not only with livelihood, but with freedom, dignity, development and participation in community.
Artificial intelligence forces us to confront what those commitments mean when human labour is no longer technologically indispensable.
There may be no human right to perform a task simply because a machine can now perform it.
There may be no human right to preserve every occupation against technological change.
And there is certainly no human-rights imperative to preserve dangerous, degrading or meaningless labour.
But if work remains one of the principal arenas in which human beings exercise agency, develop capabilities and participate in society, then the progressive removal of human beings from meaningful work cannot be treated as an economic question alone.
The challenge for the age of AI is therefore not simply to ensure that people continue to have jobs.
It is to decide what place human beings should continue to occupy within the world of work and what, ultimately, the right to work is a right to do.
Centre for Digital Ethics - September 2026
