Is AI leaving women behind?

In recent months, we have seen developments in artificial intelligence (AI) take gigantic steps forward, particularly with the arrival of coding agents. Unlike chatbots that are focused mostly on answering questions, new tools like OpenAI’s GPT-5.3 Codex and Anthropic’s Claude code can now execute tasks. AI agents can understand a goal, break it into steps, take action using online tools, and adjust the plan if something goes wrong. For instance, people can use AI agents to book flights, organize their files, conduct research and compare insights, and write and deploy code. In other words, certain white-collar roles may soon be replaced by AI.

While these tools remain imperfect and not yet fully easy to use, the technology is developing at an accelerated trajectory. As interfaces improve and access expands, adoption is also expected to rise quickly, disrupting everyday workflows and reshaping certain industries. This has raised important and urgent questions not only about the future of work but also about social protection and equity. And experts have increasingly warned about the unique and disproportionate threat that AI poses to women.

According to the International Labour Organization (ILO), around 29 percent of female-dominated occupations are at risk of having key tasks taken over by generative AI, compared to just 16 percent of male-dominated occupations. For jobs at the highest risk of automation, 16 percent of female-dominated roles fall into this category compared to only 3 percent of male-dominated ones.

This disparity is largely driven by structural inequalities in the labor market. As the ILO points out, women are overrepresented in routine, clerical, and administrative roles that are more easily automated. In contrast, men dominate fields that are less easily automated, like construction, manufacturing, and manual trades. Alarmingly, the Philippines is among the few countries cited by the ILO where more than 40 percent of women’s employment is vulnerable to AI. While AI is expected to create new opportunities in technology-driven sectors, women remain underrepresented in these fields. This suggests the potential risk that they may be displaced from existing roles without equal access to emerging ones.

Job displacement is only part of the problem. AI systems are shaped by existing societal biases, which they can amplify at scale. A study by the Berkeley Haas Center for Equity, Gender, and Leadership found that 44 percent of 133 AI systems across industries showed gender bias. These biases surface in stereotypical portrayals of men and women, less accurate medical advice for women, and more difficult loan approvals due to male-skewed training data. In hiring systems, they can lead to unfair decisions and unequal pay. As AI becomes more embedded in decision-making, it risks reinforcing inequalities that shape women’s access to opportunities and resources.

Even when access is available, women are significantly less likely than men to adopt generative AI tools, by about 25 percent across multiple contexts. Research by Harvard Business School Associate Prof. Rembrand Koning found that women’s hesitation stems from social and psychological pressures, particularly concerns about ethics and fears of being judged as less competent for relying on AI to help with their work. Since women tend to face higher social penalties for perceived lack of expertise, they are more cautious about how using AI could undermine how their abilities are perceived in professional settings.

To help women overcome this hurdle, organizations must actively foster a culture of psychological safety around AI use. Aside from investing in training for all employees, explicitly encouraging the use of AI without stigma can help create an enabling environment for women to participate in and benefit from AI fully.

Addressing AI bias requires intentional action at multiple levels, but it crucially begins with the people designing the systems. Developers must prioritize gender equality from the outset by using diverse, representative datasets and ensuring women’s perspectives and voices are considered in the design process. More women need to be present in rooms where decisions are being made. Investing in women’s participation and leadership in science, technology, engineering, and mathematics is critical to shaping more equitable technologies.

At a broader level, the United Nations is rightly pushing for “stronger global governance” that advances closer coordination among nations and promotes evidence-based decision-making. Embedding gender and human rights perspectives into AI policy and development ensures accountability, mitigates harmful bias, and distributes the benefits of AI more equitably.

AI is transforming the world and the workplace, but its impact is not gender-neutral. Without parallel changes in policy, deliberate efforts to build more diverse and gender-sensitive AI systems, and targeted initiatives to address occupational segregation and adoption gaps, AI risks not only reflecting existing inequalities but also deepening them.