Volume 26, Issue 1 (4-2026)
Social Welfare Quarterly 2026, 26(1): 9-41 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Maaboudi R, Nademi Y, Khochiany R. (2026). The Technology Paradox: Analyzing the Dual Impact of Artificial Intelligence on Well-Being Evidence from Leading Countries. Social Welfare Quarterly. 26(1), 9-41. doi:10.32598/refahj.26.100.8
URL: http://refahj.uswr.ac.ir/article-1-4512-en.html
URL: http://refahj.uswr.ac.ir/article-1-4512-en.html
Full-Text [PDF 614 kb]
(536 Downloads)
| Abstract (HTML) (1775 Views)
Full-Text: (12 Views)
Extended Abstract
Introduction
In recent years, artificial intelligence (AI) has emerged as a general-purpose technology with far-reaching implications for economic performance and social welfare. Unlike earlier waves of automation concentrated in manufacturing, AI diffuses across a wide spectrum of sectors, including finance, healthcare, public administration, and digital services, reshaping production processes, labor demand, and welfare distribution. This pervasive diffusion introduces a fundamental paradox: while AI enhances efficiency and economic growth, it may simultaneously generate adverse distributional and structural effects that undermine overall welfare.
The existing literature has predominantly focused on the productivity and growth effects of AI, often overlooking its multidimensional and distributional implications for welfare. Moreover, empirical studies frequently rely on single proxies, such as patents or publications, which inadequately capture the economic intensity of AI development. In contrast, welfare is inherently multidimensional, encompassing income, inequality, health, and social stability.
This study contributes to the literature by examining the impact of AI investment on welfare across 20 leading AI countries over the period 2017–2023, employing a composite AI investment index and the Legatum Prosperity Index. More importantly, it advances the literature by explicitly modeling the interaction channels through which AI affects welfare, namely human capital, income inequality, life expectancy, and economic growth, thereby providing a more comprehensive and structurally grounded understanding of the AI-welfare nexus.
Methodology
The analysis is grounded in a dynamic welfare framework in which AI affects welfare through multiple interrelated channels. Specifically, AI influences:
Economic growth, by increasing productivity and efficiency;
Labor markets, through skill-biased technological change and job displacement;
Income distribution, by altering returns to skills and capital;
Human capital, via complementarities and substitution effects;
Health outcomes, through improvements in medical technologies and service delivery.
To capture these complex dynamics, the study employs a dynamic panel data model estimated using the GMM approach. The specification includes:
A lagged dependent variable to account for welfare persistence,
Direct effects of AI,
Interaction terms between AI and key structural variables (human capital, inequality, life expectancy, and growth),
Time fixed effects to control for global shocks (e.g., post-COVID digital acceleration).
This framework allows us to disentangle the direct (efficiency-driven) and indirect (distributional and structural) effects of AI on welfare.
Findings
AI exerts a positive direct effect on welfare, reflecting its role in enhancing productivity, reducing costs, and improving access to goods and services. This confirms the efficiency-enhancing nature of AI as a general-purpose technology. However, the interaction effects reveal a different picture. The interaction between AI and human capital is negative and highly significant, indicating that existing human capital structures are insufficient to absorb AI-induced technological change. Rather than complementing labor, AI appears to substitute for certain skills, leading to skill mismatches and labor displacement. The interaction between AI and income inequality is negative, suggesting that AI disproportionately benefits high-skilled workers and capital owners, thereby exacerbating inequality and reducing welfare. The interaction between AI and economic growth is also negative, implying that AI-generated growth is not welfare-enhancing due to unequal distribution and structural disruptions. In contrast, the interaction between AI and life expectancy is positive, highlighting the role of AI in improving healthcare systems and non-material dimensions of welfare. Taken together, these results indicate that although AI generates efficiency gains, its indirect negative effects dominate, leading to a net negative impact on welfare. This finding supports the existence of a “technology-welfare paradox,” whereby technological advancement does not necessarily translate into improved societal well-being.
Discussion
The results provide strong evidence that the welfare implications of AI are fundamentally shaped by structural and distributional mechanisms rather than purely by efficiency gains. A key finding of this study is that human capital fails to mitigate the adverse welfare effects of AI. From a theoretical perspective, this can be explained by recent developments in the economics of technological change. Unlike earlier skill-biased technological change, AI increasingly exhibits characteristics of task substitution rather than task complementarity, particularly in cognitive and routine-intensive occupations. As a result, even relatively skilled workers may face displacement or wage compression. Moreover, human capital accumulation appears to lag behind the pace and direction of AI innovation. This creates a dynamic mismatch between the skills supplied by the labor force and those demanded by AI-driven production systems. Consequently, investments in human capital, in the absence of structural adaptation, may not only fail to offset AI’s negative effects but may even reinforce inequality and labor market polarization. The negative interaction between AI and economic growth further suggests that growth alone is no longer a sufficient condition for welfare improvement. When growth is driven by capital-intensive and skill-biased technologies, its benefits may be concentrated among a narrow segment of the population, reducing its overall welfare impact. At the same time, the positive effects of AI on life expectancy highlight the importance of sectoral heterogeneity. AI contributes positively to welfare in domains such as healthcare and public services, where technological improvements directly enhance quality of life. Overall, these findings imply that the welfare consequences of AI depend critically on institutional and policy frameworks. Without appropriate redistribution mechanisms, labor market policies, and adaptive education systems, AI-driven growth may lead to welfare divergence rather than welfare improvement. Accordingly, policy responses should focus on:
Aligning human capital formation with AI-driven skill demand;
Implementing inclusive growth strategies to reduce inequality;
Strengthening labor market institutions to manage technological transitions;
Promoting AI applications in sectors with direct welfare benefits, such as healthcare and education.
Finally, it should be noted that, despite the use of dynamic panel methods, the results should be interpreted as conditional empirical associations rather than strict causal effects, consistent with the limitations highlighted in the revised manuscript.
Ethical Considerations
Compliance With Ethical Guidelines
This research was conducted in accordance with ethical principles for human studies
Authors’ Contributions
All authors contributed equally to the design, execution, and writing of all parts of this research.
Funding
There is no funding support.
Conflict of Interest
The authors declare no conflicts of interest relevant to the content of this article.
Acknowledgments
We thank the anonymous reviewers for their useful comments, which greatly contributed to improving our work.
Introduction
In recent years, artificial intelligence (AI) has emerged as a general-purpose technology with far-reaching implications for economic performance and social welfare. Unlike earlier waves of automation concentrated in manufacturing, AI diffuses across a wide spectrum of sectors, including finance, healthcare, public administration, and digital services, reshaping production processes, labor demand, and welfare distribution. This pervasive diffusion introduces a fundamental paradox: while AI enhances efficiency and economic growth, it may simultaneously generate adverse distributional and structural effects that undermine overall welfare.
The existing literature has predominantly focused on the productivity and growth effects of AI, often overlooking its multidimensional and distributional implications for welfare. Moreover, empirical studies frequently rely on single proxies, such as patents or publications, which inadequately capture the economic intensity of AI development. In contrast, welfare is inherently multidimensional, encompassing income, inequality, health, and social stability.
This study contributes to the literature by examining the impact of AI investment on welfare across 20 leading AI countries over the period 2017–2023, employing a composite AI investment index and the Legatum Prosperity Index. More importantly, it advances the literature by explicitly modeling the interaction channels through which AI affects welfare, namely human capital, income inequality, life expectancy, and economic growth, thereby providing a more comprehensive and structurally grounded understanding of the AI-welfare nexus.
Methodology
The analysis is grounded in a dynamic welfare framework in which AI affects welfare through multiple interrelated channels. Specifically, AI influences:
Economic growth, by increasing productivity and efficiency;
Labor markets, through skill-biased technological change and job displacement;
Income distribution, by altering returns to skills and capital;
Human capital, via complementarities and substitution effects;
Health outcomes, through improvements in medical technologies and service delivery.
To capture these complex dynamics, the study employs a dynamic panel data model estimated using the GMM approach. The specification includes:
A lagged dependent variable to account for welfare persistence,
Direct effects of AI,
Interaction terms between AI and key structural variables (human capital, inequality, life expectancy, and growth),
Time fixed effects to control for global shocks (e.g., post-COVID digital acceleration).
This framework allows us to disentangle the direct (efficiency-driven) and indirect (distributional and structural) effects of AI on welfare.
Findings
AI exerts a positive direct effect on welfare, reflecting its role in enhancing productivity, reducing costs, and improving access to goods and services. This confirms the efficiency-enhancing nature of AI as a general-purpose technology. However, the interaction effects reveal a different picture. The interaction between AI and human capital is negative and highly significant, indicating that existing human capital structures are insufficient to absorb AI-induced technological change. Rather than complementing labor, AI appears to substitute for certain skills, leading to skill mismatches and labor displacement. The interaction between AI and income inequality is negative, suggesting that AI disproportionately benefits high-skilled workers and capital owners, thereby exacerbating inequality and reducing welfare. The interaction between AI and economic growth is also negative, implying that AI-generated growth is not welfare-enhancing due to unequal distribution and structural disruptions. In contrast, the interaction between AI and life expectancy is positive, highlighting the role of AI in improving healthcare systems and non-material dimensions of welfare. Taken together, these results indicate that although AI generates efficiency gains, its indirect negative effects dominate, leading to a net negative impact on welfare. This finding supports the existence of a “technology-welfare paradox,” whereby technological advancement does not necessarily translate into improved societal well-being.
Discussion
The results provide strong evidence that the welfare implications of AI are fundamentally shaped by structural and distributional mechanisms rather than purely by efficiency gains. A key finding of this study is that human capital fails to mitigate the adverse welfare effects of AI. From a theoretical perspective, this can be explained by recent developments in the economics of technological change. Unlike earlier skill-biased technological change, AI increasingly exhibits characteristics of task substitution rather than task complementarity, particularly in cognitive and routine-intensive occupations. As a result, even relatively skilled workers may face displacement or wage compression. Moreover, human capital accumulation appears to lag behind the pace and direction of AI innovation. This creates a dynamic mismatch between the skills supplied by the labor force and those demanded by AI-driven production systems. Consequently, investments in human capital, in the absence of structural adaptation, may not only fail to offset AI’s negative effects but may even reinforce inequality and labor market polarization. The negative interaction between AI and economic growth further suggests that growth alone is no longer a sufficient condition for welfare improvement. When growth is driven by capital-intensive and skill-biased technologies, its benefits may be concentrated among a narrow segment of the population, reducing its overall welfare impact. At the same time, the positive effects of AI on life expectancy highlight the importance of sectoral heterogeneity. AI contributes positively to welfare in domains such as healthcare and public services, where technological improvements directly enhance quality of life. Overall, these findings imply that the welfare consequences of AI depend critically on institutional and policy frameworks. Without appropriate redistribution mechanisms, labor market policies, and adaptive education systems, AI-driven growth may lead to welfare divergence rather than welfare improvement. Accordingly, policy responses should focus on:
Aligning human capital formation with AI-driven skill demand;
Implementing inclusive growth strategies to reduce inequality;
Strengthening labor market institutions to manage technological transitions;
Promoting AI applications in sectors with direct welfare benefits, such as healthcare and education.
Finally, it should be noted that, despite the use of dynamic panel methods, the results should be interpreted as conditional empirical associations rather than strict causal effects, consistent with the limitations highlighted in the revised manuscript.
Ethical Considerations
Compliance With Ethical Guidelines
This research was conducted in accordance with ethical principles for human studies
Authors’ Contributions
All authors contributed equally to the design, execution, and writing of all parts of this research.
Funding
There is no funding support.
Conflict of Interest
The authors declare no conflicts of interest relevant to the content of this article.
Acknowledgments
We thank the anonymous reviewers for their useful comments, which greatly contributed to improving our work.
Type of Study: orginal |
Received: 2025/08/28 | Accepted: 2026/05/5 | Published: 2026/05/31
Received: 2025/08/28 | Accepted: 2026/05/5 | Published: 2026/05/31
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
Contact Information
Copyright © The Author(s);
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 (CC-By-NC 4.0), which permits use, distribution, and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Contact Information
