AI and gender: deepening the divide further?

The rise of AI poses a significant risk to the job market, especially towards the marginalised communities. While AI provides a massive increase in efficiency and productivity, not everyone in the job market experiences this benefit equally. With 75% of knowledge workers now using AI, there is a constant threat to labour displacement, wage inequality and changing skills demand. AI is changing the current landscape of the labour market, and not all workers are well equipped to adapt. Occupational segregation, wage gaps and access to opportunities are now shaping how men and women are being affected by the AI revolution.

The impact of AI is predicted to vary by industry, with some sectors more vulnerable to automation than others. These include low-skilled, everyday employment such as retail, data entry and healthcare. With many such businesses dominated by women workers, the World Economic Forum emphasises that these sectors are the most vulnerable to AI automation. Women are feared to face an 11% risk of job loss due to automation, compared to 9% men. It is estimated that 26 million women's jobs in 30 countries are at high risk of displacement within 20 years, with a 70% or higher probability of automation, affecting 180 million women globally, as per a study 'On the Margins - Women Workers and the Future of Work, Narratives in Pakistan' by Friedrich-Ebert-Stiftung, a German political party foundation.

AI poses a threat to the existing wage gap between men and women. Many high-paying STEM fields are dominated by men, and these fields are likely to experience increased demand due to the rise of AI, such as jobs in machine learning, AI and engineering. A 2018 OECD report 'Bridging the Digital Gender Divide', says in most OECD countries, women are less likely to hold high levels of "digital literacy" compared to men, and this difference grows larger with age, as skill development becomes much more challenging with simultaneous employment and domestic care responsibilities.

According to ILO, AI tends to increase demand for high-skill worker while reducing demand for medium-skill worker, and completely automating low-skill jobs. This is concerning in developing economies where women majorly hold low to medium level jobs.

The use of AI in recruiting and performance evaluation has generated worries about the possibility of algorithmic bias, which might exacerbate gender disparity in the workplace. AI systems, when trained on biased data, can duplicate and even exacerbate pre-existing gender bias. Amazon, for example, discontinued their AI recruiting tool after it was discovered that the algorithm preferred male candidates by consistently lowering resumes that included the term "women".

Furthermore, AI-powered monitoring systems used for performance management may disproportionately harm women, who are more likely to take on caregiving obligations and work variable hours. If AI systems fail to account for these dynamics, they may unfairly punish women for nontraditional work habits, exacerbating the gender gap in professional development.

Another critical impact of AI is the lack of resources and skills for women to learn and adapt to AI. Women make up only 34% of the workforce in STEM fields, making it difficult for women in non-STEM fields to learn AI integrated technologies. This impedes the AI transition and potentially disadvantages women's learning.

Addressing the gendered consequences of AI necessitates focused policy initiatives and proactive approaches. First, boosting female involvement in STEM education and AI-related sectors is critical to narrowing the gender gap in the job market. Mentorship programmes, scholarships and gender-sensitive AI skills training can all help women break into rising tech positions.

Governments and enterprises must also emphasise AI accountability, ensuring that AI systems are built and deployed in a way that reduces prejudice. The EU's planned AI laws, which include stringent control of high-risk AI applications, provide a promising foundation for combating algorithmic prejudice. Furthermore, labour rules should be revised to encourage reskilling and upskilling programmes that educate workers, particularly women, for the AI-driven economy.