Women in AI: A South Asia Outlook on Representation, Equity and Empowerment

Abstract

The panel explored how women in South Asia experience the AI ecosystem—from early education to senior leadership—highlighting systemic stratifications (caste, disability, gender) that shape participation. Drawing on the newly released South Asia Outlook report and LinkedIn’s Economic Graph, speakers examined the leaky talent pipeline, the impact of automation on female‑dominated occupations, and the crucial role of universities in embedding diversity, ethics and interdisciplinary practice. The discussion concluded with concrete calls to action for skilling women, reshaping AI language and governance, and sustaining advocacy that can translate into policy change.

Detailed Summary

• The moderator opened with a series of acknowledgements and thanked the audience repeatedly, setting a courteous tone.
• An opening speaker (identified from the South Asia Outlook report team) highlighted multiple axes of stratification—caste, disability, gender—that intersect in AI systems.
• She emphasized the need for systematic audits of training data and model outcomes, stressing that representation gaps begin early (girls in STEM) and widen at each career stage, especially in governance and monitoring roles of large AI projects.

Key Insight: Women are present in STEM curricula, but their participation sharply declines in senior, decision‑making positions, creating a blind spot for identifying and mitigating AI harms.

2. Bhutan’s perspective – building an inclusive tech ecosystem

Speaker: Leki Choden (Dragon Coders, Bhutan)

• Bhutan’s geography (20 mountainous districts separated by rivers) limits early exposure to AI and robotics for girls.
• Educational gaps: Universities and schools are still incorporating AI/robotics into curricula; infrastructure upgrades and skilled faculty are slow.
• Role‑model scarcity: Few visible women in AI, making it harder for girls to envision an AI career.
• Connectivity challenges: Low‑bandwidth, expensive internet especially in rural areas hampers practical learning.
• Socio‑economic constraints: Girls from financially vulnerable families prioritize short‑term income over long‑term AI training, which requires time and investment.
• Cultural expectations: Despite Bhutan’s “Gross National Happiness” philosophy, women are still often expected to be primary homemakers, limiting their participation in tech.

Recommendation: Develop targeted outreach programs, improve internet access in remote districts, and showcase local women role models to inspire early interest.

3. LinkedIn Economic Graph – data on women in AI

Speaker: Aditi Jha (LinkedIn)

• Data scope: LinkedIn’s Economic Graph aggregates signals from 1.3 billion members (≈170 million in India) covering 40 000 skills, millions of companies, and detailed job postings.
• Top‑of‑the‑funnel problem: Across all surveyed countries, ≤ 30 % of AI talent is female.
• Leadership leaky pipeline: In professions where women represent ~50 % at entry level, only ≈ 12 % reach C‑suite positions.
• Soft‑skill advantage: Women list a higher proportion of soft skills (communication, empathy, critical thinking, creativity) than men. These are projected to be the currency of work in an AI‑augmented future.
• Automation risk: Jobs heavily held by women—customer service, paralegal work—are highly automatable, threatening disproportionate displacement.
• Opportunity: Leverage women’s soft‑skill strengths to occupy roles that augment rather than replace AI, ensuring equitable transition.

Call to Action: Increase female participation at the AI talent entry point, and design reskilling pathways that align women’s soft‑skill strengths with emerging AI‑centric roles.

4. Universities as catalysts for gender‑inclusive AI

Speaker: Vivek Menon (Amrita Vishwa Vidyapeetham University)

• Universities sit at the nexus of talent development, ethical grounding, and research.
• Data bias origin: Many AI models are trained on datasets lacking South Asian representation (e.g., Scopus data missing Indian context).
• Curriculum interventions: Embed diverse data practices and ethics modules from the earliest courses.
• Live‑in Labs programme: Multidisciplinary student‑faculty teams live in selected villages for weeks‑months, co‑creating solutions that respect local data diversity and contextual sensitivities.
• Women leadership: The university’s schools are led by numerous women, who champion inclusive AI initiatives and mentor students.
• Strategic take‑away: Gender‑inclusive AI must be designed, not retro‑fitted; universities must integrate diverse perspectives and ethical considerations from day one.

5. Audience interaction & networking

• The moderator announced that no audience questions were raised despite the allocated time.
• Attendees were invited to visit the LinkedIn data booth (Booth 4.7) for deeper exploration of the Economic Graph datasets.

6. Closing remarks – advocacy, language, and leadership

Speaker: Kanta Singh (UN Women, moderator)

• Unified aspirations: Rural and urban women share the same career dreams; the narrative should shift from “rural vs urban women” to “women everywhere.”
• Skill‑matching: Emphasized the need to skill women for AI‑resilient jobs and ensure they are placed in roles less likely to be automated.
• AI language bias: Stressed who defines AI vocabulary and who benefits; women often omit AI‑related skills from résumés, reducing algorithmic visibility.
• Leadership & governance: Highlighted a recent government order addressing non‑consensual imagery online—a policy win influenced by sustained advocacy from women‑focused tech forums.
• Call to engagement: Urged participants to enter every conference room, speak up, and ensure women’s data is present in AI systems; otherwise, women will be excluded from future decision‑making.
• Acknowledgements: Thanked partners—LinkedIn, The Quantum Hub, UNESCO, and the broader Women for Ethical AI for South Asia coalition—for enabling the conversation.

Strategic Recommendation: Maintain persistent, multi‑venue advocacy to shape AI policy, and ensure women’s professional profiles accurately reflect AI competencies to improve algorithmic matching.

7. Closing ceremony

• Brief ceremonial distribution of mementos to panelists.
• Reiterated gratitude to all participants, moderators, and partner organisations.

Key Takeaways

• Leaky Pipeline: Only ~30 % of AI talent in South Asia is female at entry level; this drops to ~12 % at senior leadership, creating a blind spot for AI bias detection.
• Soft‑Skill Strength: Women consistently list higher soft‑skill scores (communication, empathy, creativity), which are projected to be most valuable in an AI‑augmented workforce.
• Automation Threat: Female‑dominated occupations (customer service, paralegals) are highly susceptible to automation, demanding proactive reskilling pathways.
• Bhutan Challenges: Geographic isolation, limited internet, lack of AI curricula, and scarcity of role models impede early AI exposure for girls.
• University Role: Academic institutions must embed diversity‑aware data practices, ethics, and interdisciplinary “live‑in labs” from the start, leveraging women leaders to model inclusive AI development.
• Language & Visibility: Women often omit AI‑related skills from professional profiles, reducing their visibility to AI‑driven hiring algorithms; conscious self‑branding is essential.
• Advocacy Impact: Persistent advocacy by women‑focused coalitions can precipitate policy actions (e.g., government order on non‑consensual images).
• Unified Aspirations: Rural and urban women share identical ambitions; interventions should treat women as a single demographic rather than a split one.
• Actionable Pathways:
  1. Strengthen early AI/robotics exposure in schools, especially in remote districts.
  2. Boost internet infrastructure and affordable connectivity.
  3. Showcase local women role models in AI.
  4. Align university curricula with ethical, diverse‑data AI training.
  5. Encourage women to list AI competencies on professional platforms.
  6. Foster interdisciplinary, community‑embedded projects to ensure contextual relevance.
• Collaboration is Key: Ongoing partnerships among industry (LinkedIn), academia (Amrita University), NGOs (UN Women, UNESCO), and regional innovators are vital to achieve gender‑inclusive, people‑centered AI across South Asia.

---

See Also:

• empowering-the-human-edge-building-a-future-ready-workforce-in-the-age-of-ai
• flipping-the-script-how-the-global-majority-can-recode-the-ai-economy
• safe-and-trusted-agentic-ai-building-accountability-and-inclusion-for-india-and-the-global-south
• from-promising-pilots-to-system-shifts-what-it-really-takes-to-scale-responsible-ai-in-education
• evaluations-and-open-source-software-for-ai-for-social-good-at-scale
• keynote-i-to-the-power-of-ai-an-8-year-old-on-aspiring-india-impacting-the-world
• towards-a-safer-south-launch-of-the-global-south-network-on-ai-safety-and-evaluation
• policies-for-social-and-economic-resilience-in-the-ai-age-global-south-perspectives
• preparing-to-monitor-the-impacts-of-agents-closing-the-global-assurance-divide-for-safe-and-trusted-ai
• trusted-ai-at-scale-a-global-south-leadership-dialogue