India AI Impact Summit Insights
All sessions

Women in AI: A South Asia Perspective on Equity and Leadership

Contents

Executive Summary

This session from the India AI Impact Summit 2026 examines the severe gender representation gap in AI across South Asia, presenting data showing women comprise only 30% of AI engineering talent in India, dropping to 20% in Nepal and 15% in Bangladesh. Multiple speakers argue that without deliberate intervention, AI systems will embed and amplify existing gender inequalities, and emphasize that achieving equitable AI requires addressing data gaps, leadership representation, job disruption risks, and educational access across the region.

Key Takeaways

  1. The data representation crisis is a governance crisis: Until women are represented in data collection, training data, and AI system design, algorithmic systems will perpetuate and amplify gender inequality. This is not a diversity issue—it is a technical and policy imperative.

  2. Women are not a homogeneous group: Solutions must account for intersectional differences (caste, class, disability, rural/urban location, access to infrastructure). Educational access and job transition support require context-specific, locally-informed approaches.

  3. Leadership representation determines whose interests get served: Women in AI engineering roles, AI governance, and policy-making are essential—not just for equity, but because women bring different perspectives on risk, safety, and social impact. Without them, harms will be designed in.

  4. Soft skills are a strategic asset in AI economies: As routine tasks automate, human capabilities like empathy, critical thinking, and communication become the competitive advantage. Women's documented strength in these areas can be leveraged if training and hiring practices evolve accordingly.

  5. Speak up everywhere, measure progress: The session closes with urgent messaging that gender must move from "side events" to every room and conversation. Absence from data means absence from systems designed on that data—making visibility and voice urgent for both current and future generations.

Key Topics Covered

  • Gender representation disparities in AI engineering, research, and leadership across South Asia
  • Gender and AI Outlook Study findings for six South Asian countries (Bangladesh, Bhutan, India, Maldives, Nepal, Sri Lanka)
  • Data systems and feminist political economy: how unrepresentative data reproduces gender inequalities
  • AI's impact on future jobs: disruption vs. augmentation, with disproportionate effects on women workers
  • Skills gap and transition challenges: women concentrated in jobs vulnerable to automation
  • Educational barriers to AI participation, particularly in rural and mountainous regions
  • Leadership pipeline problem: women present in STEM but severely underrepresented in AI governance and decision-making roles
  • Data protection and algorithmic bias from a gendered perspective
  • Role of academic institutions in addressing bias and building diverse talent pipelines
  • Policy and governance implications of gender-blind AI systems

Key Points & Insights

  1. Representation Crisis Across Lifecycle: Women constitute only ~30% of AI engineering talent on LinkedIn in India; this figure drops to 20% in Nepal and 15% in Bangladesh. Only 26% of AI publications across six South Asian countries are led by women as primary/corresponding authors—though 71% include at least one female author, indicating women's invisibility in leadership roles.

  2. Data Systems Perpetuate Inequality: Unrepresentative training data and algorithmic inferences create what Anita Guru Morti describes as "a new compact between capitalism and patriarchy." Traditional data stacks (land records, RTC records) invisibilize women farmers and wage laborers, leading to gender-blind downstream innovation.

  3. Concentrated Job Disruption Risk: Women are disproportionately concentrated in jobs vulnerable to AI automation (customer service, paralegal work, administrative roles). In India, Bangladesh, and Pakistan, up to 80% of some women-dominated jobs face disruption, whereas men hold more secure, higher-paid positions.

  4. Skill Transition Inequality: Women are less likely to transition to "insulated" or secure jobs; instead, they shift between disrupted job categories. Without systemic support, the AI transition will intensify stratification rather than improve inclusion.

  5. Top-of-Funnel Problem in AI Talent: Starting at only 30% female representation in current AI talent, the pipeline problem will worsen at leadership levels. In other professions where women comprise 50% of entry-level workers, female leadership falls to 12%—a pattern likely to repeat in AI.

  6. Soft Skills as Strategic Advantage: Women demonstrate stronger soft skills (communication, leadership, empathy, critical thinking) on LinkedIn profiles. These human-centered skills will become increasingly valuable in an AI-augmented workforce, presenting an opportunity if women can access training and positioning.

  7. Infrastructural and Cultural Barriers in Smaller Markets: In Bhutan, challenges include geographic isolation (20 districts separated by mountains/rivers reducing girl exposure to AI programs), inadequate internet infrastructure, lack of visible role models, cultural prioritization of women as homemakers, and economic pressure on vulnerable families.

  8. Data Protection Gaps: South Asian data protection frameworks lack adequate gendered definitions of sensitive personal data, particularly in health domains. This enables algorithmic discrimination in credit decisions, ad targeting, and service access.

  9. Invisible Labor in Global Value Chains: Data workers, content moderators, and gig workers—disproportionately women—remain invisible in traditional labor statistics, making it impossible to assess their vulnerability to AI disruption or design protective policies.

  10. Critical Juncture for Governance: The panel emphasizes this is the "first floor/ground floor" of AI governance discussions. Unlike the internet era, where gender was ignored until embedded in systems, there is now an opportunity to center women in AI design, policy, and governance—but this requires action across all sectors immediately.

Notable Quotes or Statements

  • Tim Curtis (UNESCO): "Without deliberate and informed interventions and efforts, the systems we build today risk embedding and amplifying longstanding gender inequalities."

  • Anita Guru Morti (IT for Change): "The situation in the data and AI economy is aiding a new kind of compact, a new kind of contract between capitalism and patriarchy."

  • Emad Karim (UN Women): "Once upon a time in 2023, generative AI boomed... we are seeing the same stories. It is happening again... Gender equality should not be the job of just women and girls or just nonprofit. It should be the job of everyone, every man and woman, every gender, every policy maker."

  • Aditi Namo (independent policy expert): "While governments are now picking up on AI governance as an efficiency tool... it will improve inclusion by the nature of it but it will intensify stratification."

  • Aditi Jha (LinkedIn): "If you look at jobs that are going to get disrupted versus those that are insulated and augmented, women tend to hold a lot more of those jobs which are going to be disrupted."

  • Closing remarks: "If you are not part of the data that is being generated now you are going to be out of the data which is going to impact our lives and the generations ahead."

Speakers & Organizations Mentioned

Speakers:

  • Tim Curtis, Director and Representative, UNESCO Regional Office for South Asia
  • Yuson Kim, Chief of Social and Human Sciences Sector, UNESCO Regional Office for South Asia
  • Emad Karim, Advocacy and Innovation Coordinator, UN Women Regional Office Asia and Pacific
  • Anita Guru Morti, Executive Director, IT for Change
  • Aditi Namo, Independent Data and AI Policy Expert (background: World Economic Forum, J-PAL, Reserve Bank)
  • Leaky Chan, CEO, Dragon Coders (Bhutan)
  • Aditi Jha, Board Member and Head of Legal and Government Affairs, LinkedIn
  • Dr. Menan, Professor, School of Computing, Amrita University
  • Arajata Bharti, Founding Partner, TQatch (moderator)

Organizations & Initiatives:

  • UNESCO Regional Office for South Asia
  • UN Women, South Asia Chapter
  • Women for Ethical AI Network (global and South Asia chapter)
  • LinkedIn
  • Quantum Hub
  • Amrita University
  • IT for Change
  • Dragon Coders (Bhutan)
  • Government of India
  • J-PAL (Abdul Latif Jameel Poverty Action Lab)

Technical Concepts & Resources

Datasets & Methodologies:

  • Gender and AI Outlook Study (South Asia): UNESCO's forthcoming regional report mapping women across six countries (Bangladesh, Bhutan, India, Maldives, Nepal, Sri Lanka) in education, research, workforce participation, and digital access
  • LinkedIn Economic Graph: Dataset of 1.3 billion members, ~170 million in India, 40,000+ skills, company data, and job listings enabling real-time labor market analysis by gender and demographics
  • Scopus data: Citation/publication database; note raised that global datasets fail to capture South Asian AI research contexts

Technical/Policy Concepts:

  • Data stacks: Aggregated traditional statistical systems (land records, RTC/registry records) entering digital governance systems
  • Algorithmic inference: Systems making inferences about identity characteristics (gender, caste, economic status) from limited data, enabling discrimination in credit, advertising, and service access
  • Job categorization (LinkedIn):
    • Insulated by AI: Low automation risk; require human interaction (farming, nursing, medicine)
    • Disrupted by AI: High automation risk; easily replaceable (software engineering, customer service, paralegal work)
    • Augmented by AI: Mixed human-AI collaboration
  • Soft skills (emerging as critical in AI-forward economies): Communication, leadership, empathy, critical thinking, creativity

Policy Frameworks Mentioned:

  • IT Rules (India, recently revised to address non-consensual intimate imagery)
  • Data protection acts across South Asia (noted as inadequate on gendered dimensions)
  • Gross National Happiness (Bhutan's development philosophy)

Initiatives/Programs:

  • Living Labs Program (Amrita University): Multidisciplinary student-faculty teams embedded in villages to solve contextual problems while factoring data diversity and local constraints

Additional Context

The session was framed as the inaugural regional dialogue on women in AI for South Asia at the India AI Impact Summit 2026, organized by UNESCO, UN Women, LinkedIn, the Women for Ethical AI network, and academic/tech partners. The urgency was grounded in the observation that AI is moving from a "distant technological frontier" to a structuring force in economies, institutions, and daily life—making gender equity in AI systems a critical governance issue, not a peripheral concern.