Context

• The rapid rise of Artificial Intelligence (AI) has sparked debates about its potential to transform industries, economies, and societies at large.
• Geoffrey Hinton, Nobel Laureate and pioneer of AI, recently warned that this revolution may enrich a select few while leaving the majority poorer.
• His cautionary note echoes a historical parallel known as the Engels’ pause, a period during the early Industrial Revolution when economic output soared but living standards for ordinary workers stagnated.
• The question that looms today is whether the world is entering a modern Engels’ pause in the AI era, one where productivity accelerates but shared prosperity lags behind.

The Engels’ Pause in Historical Context

• Coined by economist Robert Allen, the term Engels’ pause refers to early 19th-century Britain, where despite immense industrial expansion, real wages stagnated and inequality deepened.
• While Britain became the workshop of the world, ordinary households saw little improvement in welfare, as food absorbed most of their income and social inequalities widened.
• It was only decades later, through reforms, institutional changes, and new social contracts, that broad-based prosperity emerged.
• This historical paradox frames a crucial debate in the present: whether AI, as a general-purpose technology (GPT) akin to steam power or electricity, will generate similar delays between technological progress and widespread welfare gains.

AI as a General-Purpose Technology

• AI fits the profile of a GPT, capable of reshaping multiple industries simultaneously.
• As Agrawal, Gans, and Goldfarb (2018) argued, its unique economic contribution lies in drastically lowering the cost of prediction.
• Yet, as history shows, GPTs bring both growth and dislocation. They demand complementary innovations, institutional adaptation, and new skills before benefits trickle down.
• Without such adjustments, the gains are often captured by a few entrepreneurs or dominant firms, leaving the broader workforce vulnerable to job displacement and wage stagnation.

Signs of a Modern Engels’ Pause

• Productivity without wage growth
  • In Philippine call centres, AI copilots have boosted productivity by 30–50%, improving efficiency and cutting costs for firms.
  • Yet, workers’ wages have remained stagnant, and workloads have intensified. Rising living costs only compound the sense of declining welfare.
• Rising costs of complements
  • To remain relevant in the AI economy, workers must constantly reskill through coding boot camps, certifications, and training programs.
  • These costs parallel the 19th-century phenomenon where higher nominal wages were offset by surging food prices, leaving workers no better off.
• Concentration of gains and inequality
  • PwC projects that AI could add $15.7 trillion to global GDP by 2030. However, most gains are likely to accrue to the U.S., China, and a handful of firms controlling foundational models.
  • For much of the global workforce, particularly in developing economies, welfare improvements may be delayed or even denied.
• Job displacement and task transformation
  • From AI-powered hospitals in China to AI adoption in airports and public administration, tasks are rapidly being transformed.
  • While some roles are complemented, others are displaced, raising concerns about structural unemployment and growing inequality.

Lessons from History and Policy Responses

• The Industrial Revolution eventually delivered prosperity, but only after decades of inequality and social unrest, which prompted reforms such as trade unions, public education, and welfare institutions.
• The same logic applies today: governance, not technology alone, determines whether AI delivers broad human welfare.
• Key policy responses include:
  • Skills transition programs: Initiatives like Singapore’s SkillsFuture or Abu Dhabi’s AI-focused university (MBZUAI) exemplify efforts to build human capital for an AI-driven future.
  • Redistribution of AI rents: Mechanisms such as robot taxes, Universal Basic Income (UBI), and philanthropic contributions could help spread the benefits of AI more equitably.
  • AI infrastructure as a public good: Compute power and data access, the lifeblood of AI, must not remain prohibitively expensive or monopolised. Publicly funded open AI models, such as those launched in the UAE and Switzerland, mark steps in this direction.

Counterarguments and Optimism

• Critics of the Engels’ pause analogy argue that contemporary societies are better equipped than 19th-century Britain.
• Stronger welfare systems, faster technological diffusion, and the potential for AI to lower costs in healthcare, education, and energy could shorten or even prevent an extended pause.
• For instance, smartphones reached billions in under a decade, and AI assistants could follow a similar trajectory if deployed equitably.
• However, optimism must be tempered by caution. While macroeconomic gains may appear, many individuals may still face stagnant wages, rising costs, and insecure livelihoods.
• The challenge lies in ensuring that AI becomes a human welfare revolution, not just a productivity revolution.

Conclusion

• The spectre of a modern Engels’ pause warns that technological progress does not automatically translate into human welfare.
• History underscores that political will, institutional adaptation, and inclusive governance are crucial to bridging the gap between productivity and prosperity.
• AI governance today faces a stark choice: whether to allow an era of concentrated wealth and stagnant welfare, or to craft policies that ensure AI becomes a driver of broad-based human progress.
• The outcome, as history reminds us, will not be determined by technology alone, but by collective political and social choices.