The Abdication of Competence

Work is not simply a mechanism by which goods are produced and wages distributed. Meaningful work disciplines the individual, creates obligations towards others, transmits standards of excellence and gives a person a place within a web of reciprocal dependence. A society which regards human labour only as an inefficiency to be removed therefore misunderstands what work has historically done for civilisation.

The danger posed by artificial intelligence is consequently greater than unemployment. When difficult intellectual and creative work is delegated to machines, people lose the repeated struggle through which judgement, intuition and mastery are acquired. The carpenter does not merely produce the cabinet; in producing it repeatedly he becomes a carpenter. The same is true of the programmer, architect, engineer, writer and physician. Removing the labour may also remove the formation of the person.

This creates an epistemic problem. AI allows inexperienced people to produce outputs carrying many of the surface characteristics of expertise without acquiring the knowledge from which expertise ordinarily emerges. The distinction between knowing something and knowing how to generate a plausible answer therefore begins to collapse. As fewer people acquire deep competence, society may retain enormous quantities of information while losing the human capacity required to distinguish truth from sophistication, quality from imitation and understanding from confidence.

Artificial intelligence simultaneously reduces the cost of producing text, imagery, software, designs and commercial propositions towards almost nothing. Human attention, money and material resources do not expand at the same rate. The result is likely to be an enormous increase in the quantity of things competing for a finite capacity to notice and evaluate them. As production becomes abundant, the commercially valuable skill increasingly becomes the ability to determine what captures attention rather than what possesses enduring value.

That transition creates a darker incentive. Once AI can model individuals continuously, commercial optimisation can move beyond conventional advertising towards behavioural optimisation at industrial scale. Personalised persuasion is already more effective in some experimental settings, while existing digital markets employ interface designs that exploit behavioural biases to increase attention, purchases and disclosure of information. AI makes it conceivable that these techniques become individually generated, continuously tested and refined against each person’s particular vulnerabilities. The danger is not necessarily literal pharmacological drugging, but an economy increasingly capable of constructing addiction like systems around personalised reinforcement, stimulation and compulsion (Matz et al., 2024; Salvi et al., 2025).

Nothing within a market mechanism itself determines that this optimisation ought to stop. Markets communicate demand and coordinate exchange, but they do not independently provide a moral account of which desires ought to be cultivated. Where retention, consumption or behavioural dependence increases revenue, competitive pressure rewards firms capable of producing it unless restrained by law, professional ethics, cultural norms or institutions. Governments may attempt such restraint, but political institutions necessarily deliberate through evidence, procedure and legislation while digital systems can alter themselves continuously. The danger is therefore that commercial optimisation acquires a speed which public institutions cannot match. Existing research into dark commercial patterns already demonstrates how behavioural manipulation can be embedded within ordinary digital systems (Federal Trade Commission, 2022; OECD, 2022).

This reveals a deeper transformation of leisure. Industrial technology promised liberation from physical hardship by reducing the quantity of work required for survival. Yet the time released from labour did not simply become contemplation, family life, civic participation or genuine leisure. Much of it became incorporated into another market. Attention itself became commercially valuable. Watching, scrolling, clicking, searching, posting and responding now generate behavioural information and commercial opportunities. We may therefore cease working in the traditional sense while still functioning as productive inputs to economic systems during what we experience as leisure.

There is an irony here. We imagined that freedom meant liberation from work, yet work had often been one of the places in which discipline, competence, mutual dependence and meaning were formed. As labour diminished, leisure expanded, but leisure itself was colonised by markets competing for attention. The individual who appeared increasingly free became increasingly measurable, predictable and commercially useful. Hardship was reduced, but the freedom supposedly created by its removal was frequently redirected into consumption.

This helps explain why convenience has been such a powerful agent of technological change. Human beings naturally prefer an immediate reduction in effort when the long term consequences are distant and distributed. A machine which saves an hour presents its benefit immediately; the skill that disappears because nobody practises it may not be noticed for a generation. Each individual decision can therefore be rational while their accumulation produces a society less capable of acting without the systems upon which it has become dependent.

It would nevertheless be wrong to imagine that no alternative traditions have existed. The Arts and Crafts movement, distributism, cooperative ownership, localism, smallholder traditions and various movements concerned with craftsmanship, conservation and appropriate technology have repeatedly attempted to resist the reduction of human life to industrial efficiency. Their difficulty is structural. They usually demand higher immediate costs, slower production, greater personal responsibility and voluntary restraint in exchange for goods whose value is dispersed across community, beauty, resilience and future generations. Industrial systems offer convenience immediately while many of their losses become visible only after the practices they displaced have disappeared.

This is where Roger Scruton’s defence of inherited institutions becomes essential. Institutions, customs, professions, families, churches, associations, ceremonies and local forms of belonging are not merely administrative arrangements. They are repositories of accumulated human knowledge. Their intelligence is often embodied rather than articulated: contained in habits of conduct, expectations of reciprocity, standards of judgement, manners, rituals and forms of authority learned through participation. Much of this knowledge cannot be adequately written down because its meaning exists in the practice itself (Scruton, 2014).

A culture therefore possesses its own form of metis. Just as the craftsman acquires knowledge through repeated engagement with material, a people acquires knowledge through prolonged engagement with one another. Customs which appear arbitrary may encode settlements to conflicts whose origins have been forgotten. Ceremonies may preserve obligations which cannot easily be justified in utilitarian terms. Professional traditions may contain standards of restraint that no regulation could fully specify. Manners may embody an understanding of how strangers can occupy the same social world without perpetual confrontation. Their rationality often becomes visible only after they disappear.

The intellectual movements which came to dominate significant parts of Western thought during the twentieth century increasingly approached such inheritances through suspicion. Institutions were interpreted through categories of domination, ideology, exclusion, hierarchy and power. Some of those criticisms exposed genuine injustices and necessary reforms. The deeper problem arose when suspicion ceased to be a corrective and became a governing method of understanding society. An institution no longer needed to be understood before it could be condemned; its inheritance could itself become evidence against it (Scruton, 2015).

Scruton’s warning was that a civilisation cannot endlessly deconstruct the structures through which its members inherit belonging without eventually destroying knowledge it does not know how to reproduce. An institution can be dismantled in an afternoon while the tacit social knowledge embodied within it may have required centuries to form. Once such knowledge has disappeared, discovering that the institution served purposes beyond those visible to its critics does not restore it. Culture, like craftsmanship, depends upon transmission.

The result is that metis has not been lost only from material production. It has been lost from culture itself. We have become increasingly capable of analysing traditions as systems while becoming less capable of inhabiting them. The knowledge of how to belong to a community, how to inherit obligations, how to maintain a household across generations, how to exercise authority without tyranny, how to accept duties one did not personally choose, and how to understand oneself as the recipient rather than the author of a civilisation cannot simply be recovered from books once the practices carrying those forms of knowledge have ceased.

This cultural loss and the technological loss of practical knowledge are therefore manifestations of the same error. Both assume that explicit information can substitute for embodied knowledge. The modern state imagines that local practice can be replaced by administration; the market imagines that craft can be replaced by production; the ideologue imagines that inherited culture can be reconstructed according to theory; and artificial intelligence now tempts us to imagine that thought can be replaced by access to answers. In each case something difficult to measure is treated as expendable because its value cannot easily be represented within the system attempting to replace it.

James C. Scott’s concept of metis gives this problem a broader language. Metis is situated practical knowledge acquired through experience and adaptation. It exists in the farmer who understands a particular field, the mason who reads stone, the engineer who hears that a machine is running incorrectly and the community that knows why an apparently peculiar custom prevents recurring conflict. Such knowledge cannot simply be downloaded after the practitioners disappear because part of the knowledge was never explicitly recorded (Scott, 1998).

Technological globalisation extends the same process geographically. Containerisation, cheap energy, telecommunications, financial integration and highly specialised manufacturing allowed production to be divided across the planet with extraordinary efficiency. Nations consequently ceased producing many things which could be obtained more cheaply elsewhere. The result was abundance, but also dependence: countries became connected not merely through trade but through productive chains which few individual nations could reproduce independently (Frémont, 2009).

This created nations of products and, eventually, nations of reliance. A country may possess supermarkets filled with food while depending upon international fertiliser, fuel, machinery, logistics, refrigeration, packaging and finance to keep those shelves full. It may possess advanced manufactured goods while lacking the industrial ecology required to make their constituent parts. Apparent abundance can therefore coexist with profound productive fragility. The United Kingdom’s own food security assessments acknowledge both significant reliance on imported food and the dependence of domestic production upon global inputs and supply chains (DEFRA, 2024; DEFRA, 2026).

The relationship between technology and metis makes that dependency more severe. As technological systems become easier to use, fewer people need to understand the processes beneath them. Skills which once existed diffusely throughout a population become concentrated inside increasingly specialised organisations. Over generations, the machinery remains while the distributed human knowledge from which independent production could once be reconstructed disappears.

Dependence then becomes deeper than reliance upon a particular machine. Society becomes dependent upon the entire technical civilisation capable of manufacturing, powering, repairing and coordinating that machine. If those systems fail, possessing the final device is of little value when the productive knowledge surrounding it has disappeared.

Artificial intelligence threatens to carry this transformation from physical production into cognition itself. Industrialisation separated consumers from the knowledge required to produce many of their material necessities. AI may separate people from the intellectual processes required to produce knowledge. We may eventually possess unprecedented access to answers while becoming less capable of reconstructing the reasoning by which those answers ought to be judged.

Nor should technological abundance be confused with decentralisation. Frontier artificial intelligence depends upon immense concentrations of computing infrastructure, electrical power, specialised hardware, capital, data and technical expertise. If machine intelligence becomes one of civilisation’s principal productive resources, control of the strongest systems could concentrate extraordinary economic and political power. The individual may receive unprecedented capability through a cheap interface while becoming increasingly dependent upon institutions over which he possesses almost no control.

The resulting danger is therefore not merely technological unemployment. It is a civilisation in which production increases while human competence contracts; information proliferates while judgement weakens; leisure expands while attention is captured; global abundance grows while local productive capacity disappears; inherited institutions decline while the systems replacing them become increasingly centralised; and individuals experience themselves as autonomous precisely as their dependence upon distant technical structures becomes almost absolute.

The central political question of the AI age is therefore not simply what machines should be permitted to do. It is which human activities must remain human even when machines can perform them more cheaply, rapidly and accurately. Education, craftsmanship, mathematics, programming, agriculture, construction, care, memory, artistic practice and participation in local institutions may require deliberate protection because their value lies partly in what performing them makes of the performer.

A humane technological civilisation would therefore require boundaries. Here the machine serves. Here the human practises. Here efficiency is deliberately refused. Such restraint will appear economically irrational wherever the market value of an activity is confused with its human value. Yet without these protected domains, technological progress may ultimately succeed in relieving human beings not merely of hardship, but of the very practices through which they become capable of freedom.