When AI gets smarter, who gets paid?

 If artificial intelligence produces more, but people earn less, the question is no longer about innovation.

From Liew Li Xuan

What happens when intelligence becomes abundant, but income begins to disappear?

A recent thought experiment by Citrini Research titled “The 2028 Global Intelligence Crisis” imagines a future where artificial intelligence (AI) not only enhances productivity but also erodes the foundations of the economic system itself.

In this scenario, firms increasingly replace highly skilled workers with advanced AI systems. Output continues to grow, yet human income steadily declines. Economies keep producing, but fewer people are able to participate meaningfully in consumption.

At first glance, this may sound exaggerated. Technological change has always displaced some jobs while creating others. However, early signals suggest this wave may be different in both speed and scope.

Goldman Sachs estimates that up to 300 million jobs globally could be affected by AI, while McKinsey & Company projects that as much as 30% of current work hours could be automated within this decade.

Firms are not only experimenting with AI but beginning to restructure hiring around it. IBM, for instance, has already indicated a pause in recruitment for roles that can be replaced by automation.

The implications are no longer confined to theory. Following the circulation of these ideas, global markets have shown heightened sensitivity to AI narratives. Technology stocks have surged on expectations of productivity gains, while concerns about labour displacement and income inequality have begun to surface in investor discussions.

This divergence reflects a deeper tension. Markets are pricing in efficiency, but not necessarily the long-term stability of demand.

This raises a critical question: if firms become more productive but workers earn less, who sustains consumption?

An economy cannot consume what its people cannot afford to buy.

For Malaysia, this question is particularly urgent. The country already faces structural challenges within its labour market.

Youth unemployment remains consistently higher than the national average, and graduate underemployment continues to reflect a mismatch between education and available jobs.

At the same time, the rise of gig work has introduced new forms of income volatility, especially among young Malaysians navigating an increasingly fragmented employment landscape.

Malaysia’s social protection architecture, anchored in institutions such as the EPF and Socso was designed for stable, long term employment relationships. That assumption is already eroding.

A future in which AI reduces demand for both routine and cognitive labour may widen existing protection gaps, particularly for those in informal or platform-based work.

Public discussion on artificial intelligence in Malaysia has largely focused on innovation, investment, and competitiveness. Initiatives such as the National AI Roadmap and broader digital economy strategies signal strong ambition.

Yet these conversations remain disproportionately supply focused. Far less attention has been given to how the gains from AI will be distributed and whether existing institutions are equipped to manage large-scale labour disruption.

This points to a deeper structural risk: the decoupling of productivity from wages. If output continues to rise while labour income declines, economic growth becomes increasingly disconnected from lived reality.

Growth without income is not prosperity, it is imbalance.

Artificial intelligence should not be viewed solely as a source of risk. Its potential to enhance productivity, improve public services, and support economic development is significant.

The challenge is not whether AI should be adopted, but how it should be governed.

Malaysia will need to move beyond broad commitments to innovation and begin addressing distributional questions more directly.

This includes expanding social protection systems to reflect non-traditional forms of work, strengthening access to reskilling and lifelong learning, and exploring mechanisms to ensure that productivity gains are more widely shared.

Firms that benefit disproportionately from AI driven efficiencies may also need to contribute more directly to workforce transition efforts.

For young Malaysians, this is not simply a question of employment. It is a question of whether the social contract itself will evolve.

Education has long been seen as the pathway to upward mobility, but if AI begins to substitute the very cognitive skills that education seeks to build, this assumption may no longer hold in the same way.

The most significant risk is not that artificial intelligence advances too quickly, but that institutions and policies fail to keep pace.

By the time large scale displacement becomes visible, the economic and social costs may already be deeply embedded.

The scenario outlined by Citrini Research may not unfold in its most extreme form. But it forces a necessary shift in perspective.

The real concern is not that machines are becoming more capable, but whether society can adapt in a way that remains fair, inclusive, and sustainable.

If artificial intelligence produces more, but people earn less, the question is no longer about innovation.

It is about who gets to participate in the future economy. - FMT

Liew Li Xuan is a youth advocate and founder of LifeUp Malaysia, an organisation dedicated to digital well-being, preventing cyberbullying and promoting scam awareness. She is an FMT reader.

The views expressed are those of the writer and do not necessarily reflect those of MMKtT.