Tech Explained: Leveraging AI For A Gender-Responsive Care Economy – Analysis – Eurasia Review in Simple Terms

By Sharon Sarah Thawaney

Artificial intelligence (AI) is increasingly presented as a solution to mounting pressures on care systems worldwide. In much of the Global South, however, care systems remain shaped by chronic underinvestment, widespread informality, and heavy reliance on women’s labour. Introducing AI into this context is therefore not a neutral technological shift. Without deliberate attention to these structural conditions, AI is likely to reinforce existing inequalities rather than alleviate them.

Demographic transitions are intensifying strain on health and social care systems, as ageing populations and rising chronic disease increase demand for health and social care services. For instance, the World Health Organization (WHO) estimates a global shortfall of 11 million health workers by 2030, highlighting a widening gap between care needs and available human capacity. These pressures have accelerated AI adoption, privileging efficiency gains over considerations of informality, labour precarity, and the gendered organisation of care work.

The Gendered Labour Behind Global Care Systems

Any assessment of AI’s role in expanding care systems must begin by recognising the gendered and informal nature of care work. Globally, women spend 3.2 times more hours on unpaid care work than men, much of which occurs in informal settings. Domestic work alone employs 75.6 million people worldwide, 81 percent of whom are informally employed, with nearly 82 percent based in developing and emerging economies. In low-income contexts, unpaid care also includes essential tasks such as collecting water and fuel. A United Nations International Children’s Emergency Fund (UNICEF) study estimates that women and girls collectively spend around 200 million hours each day on water collection, often at the expense of education, health, and participation in the labour market.

Paid care work also remains highly gendered, with women comprising approximately 70 percent of the global health and social care workforce, providing essential services to an estimated five billion people, and contributing labour valued at over US$ 3 trillion annually. Yet, women are concentrated in lower-paid and frontline roles and hold only around 25 percent of leadership positions. Furthermore, analysis from the International Labour Organization–World Health Organization (ILO–WHO) global sectoral survey indicates that women face a 24 percent pay gap compared with men in these sectors.

Gendered Risks and Structural Biases in AI-Driven Care

Introducing AI into care systems demands scrutiny, not only because care work is gendered but also because the sector remains largely informal and unevenly integrated into digital economies. A common assumption holds that care work’s relational and emotional labour makes it resistant to automation. At the same time, AI can automate routine care-related tasks, particularly domestic work, with estimates suggesting that up to 40 percent of household labour could be automated within the next decade. While both claims are well established, they apply to different forms of care and carry distinct implications. For households, automating these tasks may free up time, but for paid care workers whose livelihoods depend on them, automation reduces both labour demand and income security.

The pressing concern is the digital divide, which plays a central role in shaping how AI affects care work. Women in low- and middle-income countries are 14 percent less likely than men to use mobile internet, leaving around 885 million women excluded from internet connectivity, most of them in South Asia and sub-Saharan Africa. These disparities—driven by social norms, limited education, and unpaid care burdens—determine who can access AI-complementary roles and who remains in work exposed to automation.

Platformisation is increasingly framed as a pathway to formalising care work, but emerging evidence points to more ambiguous outcomes, as seen in India, where platforms such as Uber offer domestic help on demand. While they provide flexible work, care workers often remain outside social protection, with limited pay, job security, and bargaining power, and are also at risk of algorithmic discrimination. Without safeguards, platform-mediated care can scale informality and precarity rather than enable equitable opportunities.

Beyond access, the design of AI systems can reproduce deeper structural biases. Many models are trained on labour and economic datasets that systematically undercount unpaid and informal care work. A 2024 United Nations Educational, Scientific and Cultural Organization (UNESCO) assessment of large language models found persistent gendered associations, linking women to domestic and caregiving roles and men to authority and economic leadership. When embedded in policy-facing or administrative AI tools, these patterns risk reinforcing the long-standing undervaluation of care rather than correcting it.

Governance also remains a risk, as AI-enabled care technologies operate in highly intimate settings and generate sensitive data on health, household routines, and levels of surveillance. In the absence of strong regulatory frameworks, women—particularly domestic workers, informal caregivers, and care recipients—have limited control over how this data is collected, shared, or monetised.

Designing Gender-Responsive AI for Equitable Care Economies

Realising this potential, however, depends on deliberate policy choices rather than technological adoption alone. Leveraging AI effectively in care systems requires treating the care economy as a central component of economic development. The scale of the opportunity is substantial: in India alone, the care economy is projected to exceed US$ 300 billion by 2030 and generate over 60 million jobs, while the World Economic Forum (WEF) estimates that investments of US$ 3.1 trillion in care and social jobs could generate a comparable boost to GDP and create over 10 million jobs in the United States. In this context, AI—if deployed deliberately—can reduce repetitive care burdens and support more equitable outcomes.

AI adoption in the care economy must therefore begin by addressing the structural care and skills burden borne by women. A 2025 World Economic Forum (WEF) report identifies unpaid care as one of the main reasons women cannot engage in AI-driven upskilling or reskilling. Heavy unpaid care responsibilities limit the time available for learning, meaning that AI-led transitions will exclude women unless supported by complementary policies such as subsidised childcare, flexible work arrangements, and accessible training programmes.

Bridging the gender digital divide is central to this transition. Without targeted investments in digital literacy and job-linked training for women care workers, AI adoption will continue to shift value towards digitally mediated roles. India can leverage national initiatives such as Digital India and Skill India, while states can implement context-specific training programmes to prepare women for AI-complementary roles.

As care work becomes increasingly platform-mediated, these platforms can serve as key enablers of more equitable AI adoption in the care economy by supporting digital skilling, fair task allocation, and pay transparency. This approach can ensure that AI promotes formalisation and upward mobility rather than reproducing informality.

Finally, AI design, procurement, and governance choices will determine outcomes. Tools must be evaluated for their impact on women’s paid and unpaid workloads and their potential to intensify algorithmic surveillance. Robust data governance is critical given the sensitivity of care-related information. Clear rules on consent, purpose limitation, and data minimisation are essential, with additional safeguards for children, older persons, and informal care recipients. When embedded within sustained public investment in care infrastructure, workforce protections, and strong governance, AI can support more resilient and gender-responsive care systems.

• About the author: Sharon Sarah Thawaney is the Executive Assistant to the Vice President (Development Studies), Nilanjan Ghosh, at the Observer Research Foundation.
• Source: This article was published the Observer Research Foundation.