This article first appeared in Forum, The Edge Malaysia Weekly on March 23, 2026 - March 29, 2026

Artificial intelligence (AI) is widely seen as the next engine of economic growth. Governments and businesses across the world are investing billions in AI technologies to enhance efficiency, innovation and competitiveness. Yet, a puzzling trend has emerged. Despite rapid advances in artificial intelligence, productivity growth in many economies remains modest.

This disconnect, often referred to as the AI productivity paradox, raises an important policy question — why has the surge in AI investment not yet translated into broad-based productivity gains?

AI has the potential to transform production, innovation and service delivery. However, the economic benefits of new technologies rarely appear immediately. Earlier technological revolutions, including electrification and the internet, also experienced long adjustment periods before productivity gains became fully visible.

Understanding and escaping the AI paradox is therefore crucial. The real challenge is not whether economies are adopting AI, but whether they possess the readiness and complementary drivers such as training and skill upgrading, institutional capacity and organisational change that are required to convert technological investment into sustained gross domestic product growth.

The AI productivity paradox refers to the persistent discrepancy between the rapid advancement and widespread adoption of AI technologies and the relatively modest improvements observed in macroeconomic productivity indicators. Evidence from advanced economies suggests that despite rising investment in digital technologies and automation, productivity growth has remained subdued.

In the US, labour productivity grew by only about 1.3 % annually between 2005 and 2016, significantly lower than the surge recorded during the late-1990s’ technology boom. The UK experienced a similar pattern, with productivity growth averaging just 0.5% annually between 2010 and 2022, despite rapid digital adoption.

The paradox appears universal but transitional, and its severity depends less on a country’s level of development than on its readiness to align AI adoption with the complementary assets needed to convert technological potential into real productivity gains.

Evidence shows that while some countries such as the Nordic nations, South Korea and Singapore have successfully translated digital adoption into productivity growth, others like the UK and Türkiye continue to experience significant gaps between investment and outcomes.

The AI productivity paradox is often described through Solow’s J-curve, in which at the earlier phase of AI investment, productivity may decrease or remain stagnant. Over the longer term, however, the trajectory gradually shifts towards the upward phase of the J-curve as firms and economies move beyond these transitional challenges. Once organisational restructuring is completed, workforce skills are realigned, and complementary innovations are fully integrated, AI technologies begin to deliver tangible productivity gains.

These gains manifest in improved operational efficiency, optimised processes, cost reductions, better decision-making and the emergence of entirely new business models and markets.

These trends indicate that productivity gains from AI do not occur immediately. Instead, they tend to diffuse gradually across firms and sectors, often requiring years of organisational and institutional adjustments before their full economic impact becomes visible.

Importantly, the paradox is not unique to any single country. It reflects a broader transitional phase in technological transformation. Economies that succeed in translating AI investment into productivity gains typically combine technology adoption with strong complementary capabilities such as workforce skills, digital infrastructure and effective innovation ecosystems.

Global experience therefore suggests that the productivity paradox is not a permanent condition. Rather, it represents a transitional stage during which economies gradually adjust their institutions, labour markets and production systems to fully harness new technologies.

Malaysia is entering an important phase in its digital transformation journey. Investments in digital infrastructure, intellectual property products and data-driven technologies have expanded steadily in recent years as firms and policymakers recognise the importance of AI in strengthening economic competitiveness.

Yet, productivity gains remain modest. While AI-related investments have increased, their translation into measurable improvements in labour productivity has been slower than expected. This reflects the classic characteristics of the AI productivity paradox, where technological investments expand rapidly but productivity improvements emerge only gradually.

There is also a risk that overly optimistic expectations about AI could lead to what analysts describe as a “trough of disillusionment”, a phase where early enthusiasm fades when tangible economic benefits fail to materialise.

Malaysia’s challenge therefore is not simply to promote AI adoption but to ensure that AI investments translate into productivity improvements across industries. Achieving this requires a stronger ecosystem that supports technological diffusion, workforce capability and organisational transformation.

A comparison with regional peers highlights Malaysia’s position within the AI productivity paradox.

Countries such as Singapore, South Korea and Japan demonstrate more stable trajectories where AI-related investment continues to grow while productivity improves steadily. Singapore, for example, has recorded AI investment growth of about 9.2% annually alongside productivity growth of 3.5%. South Korea recorded investment growth of 7.7% with productivity growth of 3.2%, while Japan recorded 1.1% investment growth with productivity growth of 0.4%.

Although investment still expands faster than productivity in these economies, the gap between the two is narrower and more stable. This suggests that these countries are transitioning into a more mature phase of technological adoption.

Malaysia, by contrast, appears to remain in an adjustment-intensive phase, where productivity gains lag behind capital investments in AI and related technologies. Empirical evidence based on the Report of AI Productivity Paradox in Malaysia (2025) indicates that AI investment has a positive and statistically significant impact on labour productivity, but the strength of this relationship remains weaker than in more advanced digital economies.

This suggests that while Malaysia is increasing its investment in AI, the complementary conditions required to fully convert these investments into productivity gains are still evolving.

The experience of leading digital economies provides several important lessons. Countries that have successfully translated AI investments into productivity growth share a number of common characteristics.

First, they possess strong human capital ecosystems. Education and training systems are able to rapidly supply workers with digital and AI-related skills required in modern industries.

Second, innovation capacity plays a critical role. Economies with strong research and development capabilities and effective collaboration between universities and industry are better positioned to absorb and deploy advanced technologies.

Third, effective institutions and coordinated industrial strategies help reduce the transition costs associated with technological change. These frameworks facilitate faster diffusion of technology across firms and sectors.

Finally, robust digital infrastructure and data ecosystems enable artificial intelligence to be integrated more seamlessly into production processes and service delivery.

For Malaysia, the policy implication is clear. AI investment alone may not automatically translate into productivity growth. Greater emphasis must be placed on strengthening the complementary capabilities that enable AI to generate economic value.

This includes expanding workforce reskilling programmes, strengthening digital infrastructure, enhancing innovation ecosystems and improving coordination across technology and industrial policies.

If Malaysia succeeds in aligning AI investment with skills development, innovation capability and institutional readiness, the country can escape from the AI productivity trap. More importantly, it can position itself as a competitive digital economy in Asean where artificial intelligence serves not merely as a technological trend, but as a genuine engine of productivity, growth and economic resilience.

Dr Mohamad Norjayadi Tamam is the deputy director-general of the Malaysia Productivity Corporation (MPC), where he leads the formulation and execution of strategic policies to enhance national productivity and competitiveness. This opinion piece is part of an ongoing series by The Hive, which explores how private capital drives innovation and growth in Malaysia and Asean.

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's App Store and Android's Google Play.