Beyond awareness: Closing the AI gender gap requires strategic action

By Manasi Vartak (pictured), Chief AI Architect at Cloudera

This year’s International Women’s Day theme, Give to Gain, is a reminder that investing in women’s advancement at work delivers returns for everyone. Diverse teams broaden talent pipelines, improve decision-making, and build workplaces where people are more engaged and more likely to stay.

Yet women face a double exposure risk in the Artificial Intelligence (AI) economy: underrepresentation in high-growth AI roles, and overrepresentation in functions most vulnerable to automation. According to Asian Development Bank (ADB), women in Asia Pacific represent only 23.9% of STEM researchers, below the global average of 29.3%. At the industry level, women hold only 23% of senior positions, and only 8% of senior technical roles.

This is important to address in 2026, a crucial year in the development and implementation of Agentic AI. In fact, IDC predicts that by 2027, half of enterprises will be using AI agents to redefine how humans and machines collaborate. As these systems increasingly influence business-critical decisions, organizations need to assess blind spots in who builds, tests, and governs them.

When women are missing from AI development, the impact compounds

AI systems inherit the assumptions of the environments that build them. When development teams skew toward a single demographic, bias doesn’t only show up in datasets. It can also appear in which problems are prioritized, how success is defined, which edge cases are tested, and what risks are accepted. In the agentic era, autonomy raises the stakes: small weaknesses in data, design, or oversight can be amplified once decisions are made at scale.

True inclusion means having diverse voices shape product direction and decision rights and not just representation in organizational charts. Practically, this means auditing datasets for representation gaps, testing models for unequal outcomes, stress-testing edge cases, and involving a diverse panel of human reviewers throughout the AI lifecycle.

Governance is what makes these practices consistent. Having national frameworks or guardrails in place that are built around principles of fairness, accountability, and human oversight ensures that AI is deployed responsibly, ethically,  and safely.

HR integration is critical for ethical AI deployment

According to NINEby9’s The Moment of Truth  report, only 13% of HR teams are leading key AI-related decisions, while nearly half of APAC companies reported IT primarily controls AI adoption. When HR is brought into the conversation late, workforce design decisions such as the evolution of roles, redesigning jobs, and what new skills are required, are often left unaddressed till it is too late.

This is where gender inequality deepens the rifts: neglecting workforce readiness disproportionately impacts women, and post-launch corrections to fix unintended but preventable gender imbalances are costly.

HR therefore needs to shift from a supporting role to a strategic one — ensuring reskilling, job transitions, and inclusion plans are designed from the start, rather than retrofitted once technology is already embedded. Ethical AI also cannot be outsourced to a model. It requires human judgment and accountability throughout the AI lifecycle, stress-testing edge cases, auditing datasets for representation, testing for unequal outcomes, and involving diverse reviewers throughout development and deployment.

Restructure how work gets valued in the AI era

As AI becomes embedded across core business functions, coding ability is no longer the sole marker of technical contribution. Engineers need business acumen, communication skills, and the ability to collaborate across functions because responsible AI depends on context and judgement, not just models.

This shift can create opportunity for underrepresented groups, including women, if organizations update what they recognize and reward. Programs like Cloudera’s Women Leaders in Technology (WLIT) create forums where women and allies connect, learn, and support leadership pathways. Women must see themselves reflected in leadership before that path feels accessible.

When women are given resources, opportunities, and authority in AI development, organizations gain better AI systems that work for everyone. In the agentic era, diversity in leadership and oversight should be treated as part of AI risk management.

Organizations that formalize cross-functional approaches, create transition pathways, and recognize emotional intelligence as technical capability will build better AI and advance gender equity.