How will AI affect women’s jobs?
Around 79% of working women in the US are in occupations susceptible to disruption from AI-driven automation, according to research. By comparison, just 58% of men are likely to be affected.
The study by Kenan Institute of Private Enterprise in North Carolina, paints a stark picture of how women’s overrepresentation in certain sectors intensifies the impact of AI on their jobs.
While these are predictions, in other parts of the world the impact of automation has already hit hard. Around 30,000 women tea pickers in Kenya have lost their jobs to automation in the last five years; an industry composed of 60% women.
So which jobs are the most vulnerable? And will AI create opportunities?
Gender disparity in the workforce
It’s fair to say that at the moment there are plenty of hypotheses flying around aiming to predict exactly how seismic the impact of artificial intelligence will be in the next few years. What we do know, is that there are gender divides in the workforce along sectors, level of seniority, pay, and even digital access. Women also tend to engage more in unpaid caregiving roles and less in STEM fields, contributing to a unique set of challenges. And if predictions are correct, it is lower-paid, low-skilled work that is more likely to be absorbed by AI.
Not everyone agrees, however. A UNESCO report from 2022 points out various studies that suggest occupations with predominantly male workers are in fact more exposed to the risks of automation because women are more likely to work in roles requiring soft skills. These roles, which include caregivers, doctors and teachers for example, are not so easily done by AI, and could therefore see a rise in demand.
The possible impact of AI on women’s jobs
Kenan Institute predicts that automation is 14 times more likely to impact low-wage workers, a demographic that predominantly comprises women and minorities.
A separate study by McKinsey Global Institute found that by 2030, 30% of US working hours could be automated. That’s only a little over six years away, and includes white-collar occupations like data analytics, product design, legal analysis, and research and development.
Certain industries, such as food services, customer service, and sales are expected to shrink the most due to automation and women are over-represented in these sectors. McKinsey’s research adds that Black and Hispanic workers, those without college degrees, and the youngest and oldest workers will also face challenges in finding new jobs by 2030.
AI bias in the workplace
AI only learns from the data it is fed by humans, therefore if this data is based on gendered stereotypes such as the workforce divides listed above, it is very likely to reinforce them. Consider for example the female voice given to virtual personal assistants like Siri, Alexa, and Google Nest. These add to the notion of women as carers who help out in the home – roles that are usually unpaid and unevenly distributed between women and men.
By repeating these stereotypes AI will have a detrimental effect on tomorrow’s workforce, heightening existing gender divides.
But what if the technology was used to dismantle these stereotypes and create opportunity? Fixing the bAIs, for example, is a huge bank of images that has been created using generative AI to show women in professions that historically have had a higher proportion of men. The images are available to use for free, and as likely to depict an elderly caucasian female doctor as a Muslim engineer who wears a hijab – images that can teach people from a young age that engineers and doctors are not just white men (tropes that come up repeatedly when asking gen AI tools like Dall-e for images).
Can AI create job opportunities for women?
Another initiative is FeministAI, a US-based non-profit that aims to encourage women, non-binary and LGBTQ+ people to get involved in the research and development of AI for the purpose of making it accessible, unbiased, and representative. Several of its workshops have been inspired by Dr Saffiya Noble’s groundbreaking book Algorithms of Oppression, which highlighted racism and other forms of oppression in search engines and AI back in 2018.
With currently only 29% of science R&D positions held by women worldwide, there is potential for a greater representation of women in AI design and development, as well as policy and implementation.
UNESCO’s report also points to a case study in Africa where AI could help reduce gender inequalities in agriculture. One example it cites is the Buy from Women platform, set up by UN Women. It connects women farmers to information, financing and markets using an open-source cloud-based supply chain system accessible on mobile devices. The platform uses AI to predict production levels and crop yields to help with planning and avoid distress selling.
There are a lot of unknowns, but these include opportunities. Time will tell who will seize them – big tech, or citizens wanting to survive and thrive.