The New Pharaoh: Who Governs the Future of Human Labour?
Britain is debating how to move millions of people from welfare into employment. But as artificial intelligence and automation reshape the economy, a more fundamental question is emerging: what happens when the institutions responsible for creating jobs increasingly invest in eliminating the need for them?
There is something revealing about the timing of Britain’s latest welfare debate. On November 10, 2026, Conservative Party Leader Kemi Badenoch and Shadow Work and Pensions Secretary Helen Whately presented a proposal to reduce working-age welfare expenditure by £36.1 billion annually. The plan is built around a familiar political proposition: a welfare system should provide protection for people who genuinely need it, but it should not become a permanent alternative to employment for those capable of working. It is an argument about public expenditure, individual responsibility, economic participation and the obligations citizens have towards one another.
Yet beneath this debate sits a question that deserves considerably more attention. Britain, like much of the developed world, is entering a period in which artificial intelligence, robotics and automated production are changing the relationship between economic growth and human employment. Governments are being asked to make people more employable, businesses are being encouraged to become more productive, and technology is increasingly capable of performing tasks that once required human labour. These developments are not necessarily contradictory, but they create a tension that conventional welfare policy has not adequately resolved.
The central question is no longer simply whether people should work. It is whether the economic system is being organised to provide sufficient opportunities for them to do so.
A plan built around responsibility
Badenoch’s presentation began with a political argument about preparation. Governments should know what they intend to do before entering office, she argued, rather than attempting to develop policy after assuming responsibility for the country. Her opening remarks were directed at the contrast between a political party that presents itself as having a developed programme and an incumbent government that, in her account, has struggled to establish a coherent direction.
The substance of the announcement was equally specific. The Conservatives proposed £36.1 billion in annual welfare savings, distributed across sickness and disability benefits, eligibility for foreign nationals, housing support, the two-child benefit limit, child disability payments, benefit-cap enforcement and long-term unemployment. The largest proposed saving, £13.7 billion, concerns sickness and disability benefits. A further £9.5 billion is attributed to restricting benefit and social-housing eligibility for foreign nationals with limited or indefinite leave to remain. The remaining measures address housing expenditure, child-related payments, fraud and the conditions attached to long-term unemployment.
The party’s five guiding principles are fairness, contribution, personal responsibility, welfare as a safety net rather than a lifestyle choice, and living within the country’s means. The underlying proposition is that public assistance should protect people against genuine hardship without weakening the relationship between work and economic independence.
There is a legitimate public-policy question here. Welfare systems must be financially sustainable, properly administered and capable of distinguishing between different circumstances. People with severe disabilities or illnesses may require long-term assistance, while others may need rehabilitation, training, workplace adjustments or help returning to employment. Fraud must be addressed, and governments have a responsibility to ensure that public resources are used for their intended purposes.
But the existence of a legitimate policy objective does not automatically establish that every proposed measure will achieve its stated savings, that its eligibility rules will be appropriately designed or that the wider economy can absorb the people expected to leave the welfare system. The £36.1 billion figure is the Conservative Party’s projected saving, not an independently established outcome. The practical results would depend on legislation, implementation, administrative capacity, behavioural responses and the availability of suitable employment.
That distinction matters. A savings target is not the same thing as an economic strategy.
The economy is changing while the welfare debate remains familiar
The most consequential omission in much of the political discussion is the changing nature of work itself. For generations, the relationship between economic development and employment was understood through a relatively familiar sequence. Businesses expanded, production increased, new industries emerged and workers moved between occupations as technology changed. Industrial machinery displaced certain forms of labour, but it also created new industries, occupations and markets. Governments responded through education, labour legislation, public infrastructure and social protection.
The present technological transition introduces a different set of possibilities. Artificial intelligence is moving beyond simple assistance with routine tasks. Systems can now perform increasingly sophisticated work involving research, programming, document production, analysis, customer interaction and other activities traditionally associated with professional employment. Robotics extends this transformation into manufacturing, logistics, agriculture, healthcare and physical operations.
The implications are not confined to factory workers or administrative employees. They reach into occupations that were once considered relatively protected because they depended on specialist knowledge, professional training or the ability to process information. This is not a prediction that human employment will disappear. It is a recognition that the composition of employment, the number of workers required for particular activities and the skills demanded by employers may change substantially.
The British government’s own Department for Science, Innovation and Technology published an assessment of AI capabilities and the labour market on January 28, 2026. It reported that approximately 70 per cent of UK workers are in occupations containing tasks that AI could potentially perform or enhance. Importantly, the assessment distinguishes between tasks that AI can complement and those it could potentially replace. It also acknowledges significant uncertainty about the scale, timing and causal relationship between AI adoption and employment outcomes. (GOV.UK)
That qualification is essential. Exposure to AI does not mean that an occupation will disappear. A technology may improve an employee’s productivity, change the nature of a job or create demand for new skills. Nevertheless, governments now have sufficient evidence to recognise that the relationship between technology and employment requires active monitoring and deliberate policy preparation.
The International Labour Organization’s 2025 assessment similarly estimated that one in four workers globally are in occupations with some degree of exposure to generative AI. Its central finding was that transformation of existing jobs is more likely than the wholesale elimination of occupations, while emphasising the importance of managing the transition through social dialogue. (International Labour Organization)
The distinction between transformation and elimination should not obscure the central issue. A job does not have to disappear entirely for the number of people required to perform it to decline. A business that previously required twenty employees to deliver a particular service may eventually require ten, even if its output increases. A company may expand its revenue while reducing its dependence on labour. An industry may become more productive without generating employment at the same rate as its productivity.
That is where the welfare debate encounters its structural contradiction.
The new Pharaoh
For centuries, political power has been understood through institutions, governments, monarchies, legislatures and the ownership of land and capital. Those who controlled essential resources possessed considerable influence over the economic circumstances of everyone else. Technology is now becoming another major source of productive power.
The comparison to a new Pharaoh is not intended to suggest that artificial intelligence itself has political intentions or that technology is inherently oppressive. The metaphor concerns the concentration of control. If an increasingly small number of corporations own, develop and operate the technologies that determine how goods are produced, services are delivered and economic value is generated, they may acquire an extraordinary degree of influence over the conditions of human participation in the economy.
The question is not whether these companies should be allowed to innovate. Technological progress has the potential to improve living standards, reduce dangerous work, expand access to services and increase productivity. The question is whether the economic and political institutions governing that progress are developing at a comparable pace.
A company may have a commercial incentive to replace a human process with an automated one. From the perspective of its shareholders, that decision may be entirely rational. Reduced labour costs can improve margins, increase output and strengthen competitiveness. But the consequences do not necessarily remain inside the company’s accounts.
When large numbers of workers lose employment, the costs may be distributed across households, communities, public services and government budgets. Income-tax receipts can be affected. Demand for social assistance may increase. Consumer purchasing power may weaken. Governments may find themselves supporting people whose previous employment disappeared as businesses became more efficient.
The enterprise captures some of the immediate productivity benefit, while the wider society may absorb part of the transition cost. This is not an argument that every automation decision produces those consequences. It is an argument that the consequences should be measured rather than ignored. The emerging question is one of economic governance: who determines the conditions under which technological productivity is introduced, and who is responsible for managing its consequences?
The asymmetry of accountability
There is an important asymmetry in the way society discusses individual responsibility. The person receiving public assistance is expected to demonstrate that they qualify. They may be required to attend appointments, comply with conditions, seek employment, participate in training and justify continued support. Governments can impose sanctions when those requirements are not met.
The individual is therefore treated as an economic actor with obligations towards the state and the wider community. But what equivalent expectations apply to the institutions with the greatest capacity to shape the labour market?
What obligations should accompany the ability to replace large numbers of human tasks with automated systems? What responsibility does a company have towards employees whose positions become commercially unnecessary because of its investment decisions? What should be expected of enterprises that receive public subsidies, tax incentives, government contracts or publicly funded infrastructure while introducing technologies that substantially reduce their domestic employment requirements?
These are not arguments for preventing businesses from becoming productive. Nor are they demands that every existing job must be preserved indefinitely. They concern the reciprocal nature of the social contract. A functioning economy does not exist independently of society. Businesses depend on educated populations, roads, electricity, telecommunications, public security, courts, financial institutions, consumers and the purchasing power of households. They operate within a legal and institutional framework that society collectively maintains.
The privileges and opportunities of enterprise are therefore connected to public institutions and social stability. It follows that the obligations of economic participation cannot reasonably be discussed exclusively in relation to individuals. Businesses, particularly those exercising significant economic power or benefiting from substantial public support, must also be part of the conversation. The social contract cannot be a one-directional instrument.
The politics of the visible symptom
There is another danger in the way the welfare debate is framed: the tendency to concentrate public frustration on identifiable groups while leaving the larger structural conditions insufficiently examined. This is not unique to welfare policy. Public discussions about crime, for example, can become dominated by the conduct of individuals while giving insufficient attention to the institutional and economic conditions that influence criminal activity. Accountability for an offence remains important, but understanding the circumstances that contribute to crime is equally necessary if a society intends to prevent its recurrence.
The same analytical distinction applies to welfare. Fraud exists, and those who deliberately exploit public assistance should be held accountable. Some individuals may also become disconnected from employment through prolonged inactivity, inadequate support or personal circumstances that require intervention. But these cases cannot automatically explain the entire expansion of welfare expenditure, nor can they substitute for an examination of the labour market, public health, economic restructuring and the availability of suitable work.
When a political debate becomes centred on the behaviour of welfare recipients, it risks treating the visible recipients of public expenditure as the principal explanation for a much larger economic problem. The danger is not that individual responsibility should disappear from the discussion. It is that responsibility may be applied selectively. If the state expects an unemployed person to demonstrate an active commitment to employment, it should also be prepared to examine whether employers are creating opportunities, whether education systems are preparing people for changing occupations and whether the economy is generating sufficient demand for human skills.
Otherwise, government risks making individuals accountable for circumstances that are substantially shaped by decisions elsewhere.
The missing side of welfare reform: the employer
Badenoch and Whately’s proposal places considerable emphasis on moving people towards employment. It also proposes reducing employment costs for younger workers by reversing what the Conservatives describe as Labour’s jobs tax and changing aspects of employment regulation. The stated objective is to encourage businesses to offer more first jobs, including Saturday and summer employment.
That approach recognises an important part of the equation: employers must be willing to hire. But the question is whether reducing the cost of employing people will be sufficient when businesses are simultaneously evaluating the commercial advantages of automation. For some employers, the decision may be straightforward. If an automated system can perform a task more quickly, consistently and cheaply than a human employee, the business may have a powerful incentive to adopt it. Lower employment costs may influence that calculation, but they do not eliminate the underlying commercial incentive to automate.
This is particularly important for young people entering the labour market. Entry-level employment has traditionally provided more than income. It has offered an opportunity to acquire experience, learn professional expectations, develop interpersonal skills and establish a record of participation in the economy. If businesses automate the routine tasks through which younger workers historically acquired experience, the problem becomes more complicated. The next generation may be expected to possess experience before being given the opportunity to acquire it.
Governments cannot solve that problem merely by reducing welfare payments or encouraging job applications. They must examine the supply of employment opportunities themselves.
Should governments establish limits on automation?
This is where the debate must become more ambitious. Should governments impose restrictions on how much automation a company can introduce? Should businesses be required to maintain a minimum proportion of human employees? Should there be employment quotas tied to the size of a company’s operations?
A blanket automation quota would be difficult to justify. Different industries have different operating requirements, and some forms of automation improve safety, increase production quality or make businesses viable in the first place. Requiring companies to retain unnecessary positions could reduce competitiveness, discourage investment and preserve inefficient processes. A rigid numerical restriction could also become obsolete as technology and industrial structures change.
But rejecting blanket quotas does not mean accepting an absence of regulation. Governments could develop a framework for responsible automation built around transparency, transition obligations and measurable public benefit. First, large enterprises introducing substantial automation could be required to conduct employment impact assessments. These assessments would identify the categories of work affected, the expected reduction or transformation of positions, the consequences for entry-level recruitment and the measures proposed for retraining or redeployment.
Second, businesses receiving substantial public subsidies, tax incentives or government contracts could be required to demonstrate how their investment contributes to domestic skills development, apprenticeship opportunities and economic participation. Public support should carry clearly defined public-benefit obligations.
Third, companies undertaking major workforce restructuring because of automation could be subject to enhanced retraining, redeployment and transition requirements. The objective would not be to guarantee every employee their existing position forever, but to prevent businesses from transferring the entire cost of technological change to workers and taxpayers.
Fourth, governments could explore sector-specific employment and apprenticeship commitments where they are economically justified. This would be particularly relevant in industries receiving public investment or operating through long-term government partnerships.
Finally, productivity gains should become part of a broader discussion about how economic value is distributed. Profit-sharing, employee ownership, improved wages, reduced working hours and investment in workforce development are among the mechanisms worth examining. The principle is straightforward: companies should be able to innovate, but the social consequences of innovation should not be treated as someone else’s problem.
The productivity dividend and the future of taxation
There is a fiscal dimension that makes this discussion particularly urgent. Modern welfare states depend on an economic structure in which individuals earn wages, businesses generate profits and governments collect revenue through multiple forms of taxation. Employment contributes through income taxes, payroll-related contributions and consumption. Businesses contribute through corporate taxation and other levies.
If production becomes increasingly automated while the share of economic value generated through human employment declines, governments may face pressure to reconsider how public revenue is collected. This does not automatically mean that fewer jobs will produce lower tax revenue. Higher productivity can generate larger profits, greater investment, lower prices and new economic activity. But the distribution of those gains matters.
If productivity increases substantially while income and wealth become more concentrated, the benefits of technological development may not translate into broadly shared prosperity.
A tax system designed around an earlier relationship between labour, capital and production may require reform as that relationship changes.
A direct tax on robots or every AI system is not necessarily the answer. Such proposals can be difficult to define and may discourage useful investment. The more fundamental question is whether the tax system adequately captures the economic value generated by increasingly automated production, and whether the resulting revenue supports the public institutions that make economic activity possible. The objective should not be to punish productivity. It should be to ensure that productivity remains capable of supporting the society in which it is generated.
Human agency must remain central
There is a philosophical dimension to this issue that should not be lost in the discussion of budgets, employment statistics and corporate regulation. Work is not simply an instrument for collecting wages. For many people, it provides structure, independence, social connection, professional identity and a sense of contribution. Employment allows individuals to participate in the economic life of their communities and to exercise a degree of control over their own circumstances.
That does not mean every person must be employed to possess dignity or social value. People who cannot work because of illness, disability, caring responsibilities or other circumstances are not less deserving of respect. Nor should society measure human worth exclusively through economic productivity. But meaningful access to economic participation remains important.
A society in which technology generates extraordinary wealth while large sections of the population struggle to find a productive role would face questions that cannot be resolved through welfare administration alone. The challenge is not to preserve every traditional occupation. It is to preserve the conditions under which people can participate in the prosperity of their society, even as the nature of work changes.
That may require new occupations, shorter working weeks, different education models, stronger technical training, new forms of employee ownership and a more deliberate distribution of productivity gains. It may also require governments to reconsider the relationship between employment, income and social protection. The future of work cannot be reduced to a question of whether people are sufficiently motivated to find jobs. It must also consider what kinds of jobs the economy will create, who will have access to them and how the benefits of technological progress will be shared.
A social contract for the technological age
The Conservative welfare proposal is an attempt to address a genuine question about the sustainability and purpose of public assistance. Its advocates argue that a welfare system should encourage contribution, protect those in need and reduce long-term dependency. Critics will reasonably examine the proposed eligibility restrictions, the consequences for vulnerable households and whether the projected savings can be achieved without creating costs elsewhere.
Those questions deserve serious scrutiny. But the debate should not end there. Britain is confronting two transitions simultaneously: the fiscal and demographic pressures affecting its welfare system, and the technological transformation of its productive economy. The first demands a sustainable social-protection system. The second demands a framework that ensures economic progress does not undermine the human participation on which social stability has historically depended.
Governments must therefore consider a broader settlement between citizens, employers and the state. Individuals should be expected to contribute where they reasonably can. Governments should provide education, infrastructure, opportunity and a social safety net that responds to genuine need. Enterprises should be free to innovate, but the introduction of transformative technologies should be accompanied by transparency, responsible workforce transition and an appropriate contribution to the wider economy.
This is not an argument against capitalism, automation or technological progress. It is an argument that economic progress must remain compatible with the long-term interests of the society that sustains it. The new Pharaoh does not need a crown. It may arrive in the form of a technology platform, an automated production system or a concentration of productive assets controlled by institutions with unprecedented economic reach. Its power will not necessarily come from issuing commands. It may come from determining who is needed, who is employable and who can participate in the generation of wealth.
The question for democratic governments is whether they will establish the principles and institutions necessary to govern that transition, or allow the distribution of economic opportunity to be determined almost entirely by the commercial decisions of those who own the technology. Before governments demand that people return to work, they must confront the possibility that the economy they govern is being redesigned to require fewer workers. Otherwise, they risk punishing individuals for a structural transition over which those individuals have very little control.
Comments