The study’s central question sits at the intersection of two research traditions that rarely speak to each other. On one side is the economics of artificial intelligence, a field dominated by questions of productivity, labor demand, and inequality. Landmark work by Daron Acemoglu and colleagues has shown that automation can displace workers, compress wages, and concentrate wealth, while other studies using firm-level data from Taiwan, China, and Europe find that AI adoption raises productivity and can create new, complementary tasks. On the other side is happiness economics, which since Richard Easterlin’s famous 1974 paradox has debated whether economic growth actually improves the human lot. The Delhi researchers brought these threads together by asking whether AI readiness shows up not just in gross domestic product, but in how people evaluate their own lives.

To measure subjective well-being, the authors drew on both of its classical components. The cognitive component, often called life satisfaction or life evaluation, was captured using the Cantril Ladder scores familiar from the World Happiness Report. In those surveys, respondents are asked to imagine a ladder whose top rung, scored 10, represents the best possible life and whose bottom rung, scored 0, represents the worst, and then to place themselves on that scale. National averages are built from roughly 1,000 respondents per country each year, with population-representative weights and three-year averaging used to sharpen the estimates. The affective component, reflecting day-to-day emotions and experiences, was measured alongside it, following the standard practice in the well-being literature established by psychologists such as Ed Diener.

Measuring AI itself posed a subtler challenge, because there is no single dial that reads a nation’s artificial intelligence capacity. The researchers settled on the Government AI Readiness Index produced by Oxford Insights, which assesses how prepared governments are to deploy and govern AI across dimensions such as infrastructure, skills, and institutional capacity. Using a government-focused index has a conceptual advantage for the study’s purpose: it captures policy-relevant readiness rather than private-sector hype, and it is available for a broad cross-section of countries. The 29 economies examined span an enormous development range, from Afghanistan, Bangladesh, Nepal, and Myanmar to Singapore, Japan, Australia, and the United Arab Emirates, with China, India, Indonesia, South Korea’s regional neighbors, and many others in between.

The methodological heart of the paper is its confrontation with endogeneity, the perennial headache of cross-country economics. Richer, better-governed countries are both more AI-ready and happier for many reasons that have nothing to do with AI itself, so a naive regression would overstate the causal role of technology. To address this, the authors employed instrumental-variable two-stage least squares estimation, a technique with roots in the classic work of Sargan, Basmann, Durbin, Wu, and Hausman. In the first stage, the AI readiness variable is predicted from instruments that are correlated with AI preparedness but plausibly uncorrelated with the error term driving happiness; in the second stage, the predicted values are used to estimate the effect on well-being. Specification tests drawn from the Wooldridge econometrics tradition guided the validity of the instruments, giving the positive AI-well-being association a firmer causal footing than ordinary least squares could provide.

The headline result is striking: even after controlling for a battery of macroeconomic variables, AI readiness retains a positive and statistically significant effect on subjective well-being across the 29 economies over 2019 to 2024. This is not simply a story of AI making countries richer and richer countries happier. The multivariate framework allowed the authors to probe how AI interacts with the traditional drivers of well-being, and they identified four mediating channels: economic growth, unemployment, health expenditures, and climate change. In other words, part of AI’s happiness effect flows through faster growth, part through labor-market conditions, part through expanded health spending, and part through environmental outcomes, consistent with a growing literature on AI applications in water treatment, environmental sustainability, and climate mitigation.

Each of these mediating channels carries its own tension. The unemployment channel is the most contested. Reviews of the AI-employment literature, including work by Virgelio and colleagues and econometric analyses by Mutascu, Guliyev, and Masoud, reach mixed conclusions about whether intelligent technologies destroy more jobs than they create. The OECD’s surveys of employers and workers document real anxieties about workplace AI, and studies of industrial robots in China find that public trust moderates whether automation translates into happiness at all. Yet the task-based framework of Acemoglu and Restrepo suggests that automation also creates new tasks in which human labor holds comparative advantage, and firm-level evidence from Taiwan and China shows AI-adopting firms expanding rather than contracting. The new study’s finding that AI readiness ultimately supports well-being suggests that, in the Asian context over this period, the positive channels outweighed the displacement effects.

The health channel is less ambiguous. A rapidly growing body of work documents AI’s contributions to medical diagnostics, from explainable deep-learning models for ocular disease classification to AI applications deployed during the COVID-19 pandemic and reviews of AI in mental health and positive psychology. Studies linking artificial intelligence, health expenditure, and digital financial inclusion to life expectancy reinforce the plausible pathway: AI-ready governments can deliver better health services, and better health is one of the most robust predictors of life satisfaction in the happiness literature. The climate channel cuts both ways, since AI carries a substantial energy footprint even as it offers powerful tools for optimizing energy systems, modeling climate risk, and automating environmental monitoring, a duality explored in depth by Cowls, Taddeo, Floridi, and colleagues.

What makes the study timely is the policy moment in which it lands. Governments across the Asia-Pacific, from APEC member economies drafting AI governance frameworks to national programs in India, Singapore, China, and the Gulf states, are racing to build AI capacity. The authors’ recommendation is direct: greater national preparedness for AI is expected to benefit the overall well-being of individuals in these economies, so investment in readiness is worthwhile on happiness grounds alone, not merely on competitiveness grounds. The finding also speaks to the developing-world dimension of the AI debate. Research on generative AI in developing countries suggests these technologies could accelerate development goals, and the new evidence indicates that the well-being stakes of that transition are real and measurable.

Caveats remain, as the authors themselves acknowledge through their careful framing. The country sample and time period were governed by data availability, the observation count is modest by the standards of panel econometrics, and the 2019 to 2024 window includes the COVID-19 pandemic, which disrupted both labor markets and self-reported well-being worldwide. Subjective well-being scores also vary in sample composition from year to year, and the Government AI Readiness Index measures preparedness rather than actual deployment, leaving open the question of whether realized AI adoption would show the same relationship. Still, the study marks a meaningful step: it moves the conversation about artificial intelligence beyond productivity statistics and job-loss headlines into the territory where economists have long measured what actually matters to people, and it finds that, across a diverse stretch of the Asia-Pacific, societies that prepare for AI appear to grow happier as they do so.

Subject of Research: The relationship between national artificial intelligence readiness and subjective well-being in Asia-Pacific economies

Article Title: Understanding the AI and subjective well-being relationship: the influence of macroeconomic factors in Asian economies

Article References: Khandelwal Garg, N., & Kher, G. (2026). Understanding the AI and subjective well-being relationship: the influence of macroeconomic factors in Asian economies. AI & SOCIETY. https://doi.org/10.1007/s00146-026-03347-5

Image Credits: AI Generated

DOI: 10.1007/s00146-026-03347-5

Keywords: artificial intelligence, subjective well-being, happiness economics, Asia-Pacific economies, Government AI Readiness Index, IV-2SLS regression, economic growth, unemployment, health expenditure, climate change, Cantril Ladder, endogeneity