AI and Gender Equality: How Inclusive Innovation Can Bridge Asia’s Digital Divide

Artificial intelligence (AI) is rapidly transforming economies across Asia and the Pacific, with global revenues projected by MarketsandMarkets and Investopedia to reach $407 billion by 2027. Yet the benefits of this digital revolution are unevenly shared. The United States, the People's Republic of China, and the United Kingdom dominate AI investment, while women remain underrepresented and excluded. The Asian Development Bank (ADB), drawing on studies by UNESCO, the World Economic Forum, the United Nations Development Programme (UNDP), and Boston Consulting Group, reveals that women hold only 30% of professional roles and just 12% of research positions globally. In Asia and the Pacific, only 23.9% of STEM researchers are women, far below the global average of 29.3%. Without deliberate inclusion, AI risks amplifying rather than bridging these inequalities.

Bias Built into the Machine

Gender inequality runs through every phase of AI's life cycle, from data design to governance. Bias starts with who builds AI and the assumptions they bring. Research shows that 44% of tested AI systems exhibited gender bias, and 25% showed both gender and racial bias. Large language models such as GPT-2 and Llama 2 have generated sexist content, associating women with domestic or submissive roles. The prevalence of female-voiced digital assistants reinforces stereotypes of female servitude, while the dominance of English-language datasets marginalizes women from non-English-speaking and rural communities. This imbalance limits not only fairness but also the accuracy and cultural sensitivity of AI systems.

Access, Safety, and the Digital Divide

When deployed, these biased technologies deepen existing inequities. According to the GSM Association (2023), women in South Asia are 32% less likely than men to use mobile internet, and only 20% of women in low-income countries are connected at all. Structural barriers, limited financial inclusion, care burdens, and inadequate training further hinder participation. Women often lack the time and trust to develop digital skills, while privacy violations and technology-facilitated gender-based violence (TFGBV) discourage their online engagement. As a result, women risk becoming passive users rather than active shapers of AI. These gaps are not confined to access but extend to employment and opportunity: hiring algorithms have penalized résumés containing "women's," and credit models disadvantage those with informal or part-time work.

Inclusive AI in Action

Despite these challenges, examples across Asia show how gender-inclusive AI can drive empowerment. In India, Karya, a social enterprise supported by Microsoft, Google, and the Bill & Melinda Gates Foundation, employs over 30,000 rural women to create inclusive datasets in local languages, ensuring that roles like "doctor" are not automatically coded as male. By paying fair wages and designing offline-friendly tools, Karya enhances both algorithmic fairness and women's livelihoods. In Indonesia, Amartha, a peer-to-peer lending platform supported by Accion and the Mastercard Center for Inclusive Growth, uses AI-based credit scoring to expand access for women entrepreneurs. The system tripled eligible borrowers, cut loan approval time to under 24 hours, and removed patriarchal barriers such as requiring a husband's signature. Meanwhile, India's Safecity platform uses AI and crowdsourced data to map unsafe areas for women, helping inform better policing and urban safety measures. In governance, the Republic of Korea's response to digital sexual crimes after the 2020 "Nth Room" scandal illustrates how adaptive legislation, expanding definitions of digital crimes, increasing penalties, and holding internet platforms accountable, can make digital spaces safer for women.

A Road Map for an Equal Digital Future

The ADB brief sets out four priorities for building gender-responsive AI. First, policy and governance: governments must embed gender equality into national AI strategies through sex-disaggregated data, gender audits, and inclusion in oversight bodies. Second, programs and projects: AI tools should be codesigned with women and evaluated by empowerment outcomes rather than access rates. Third, finance and innovation: investors should expand funding for women-led and gender-inclusive AI initiatives that demonstrate measurable social impact. Finally, capacity development: countries must strengthen women's digital literacy, STEM education, and leadership pipelines to ensure they become creators, not just consumers, of technology.

ADB's regional assessment shows how far most countries still need to go. Of the 14 Asian nations with AI strategies, only Malaysia and Pakistan explicitly integrate gender equality into policy frameworks. Others mention diversity only in passing or under ethics and inclusion clauses, leaving gender considerations largely absent from implementation.

The brief concludes that gender equality must be a core principle of AI governance. Governments, the private sector, and civil society must collaborate to ensure that AI reflects the realities and aspirations of all users. By investing in women's representation, literacy, and leadership, and embedding equality into every phase of AI development, Asia and the Pacific can transform technology from a mirror of inequality into a tool for justice and opportunity. In doing so, the region can ensure that digital transformation becomes not just a story of innovation, but one of shared progress and empowerment for all.