Opportunities and threats posed to women by artificial intelligence
At SD Law, every August we take the opportunity to reflect on the challenges faced by women in South Africa, the country with one of the highest rates of gender-based violence (GBV) in the world. We also celebrate, in the spirit of Women’s Month, the achievements of women in our society and our progress towards a fairer and more equal society…for women and for all communities in our diverse nation. This year, we are particularly inspired – and at the same time concerned – by the relationship between women and artificial intelligence (AI).
A lot has changed in a year. This time last year, no one had heard of ChatGPT. (It launched on 30 November, 2022.) We vaguely knew about artificial intelligence and machine learning but, aside from our Netflix recommendations, we probably had little awareness of it. AI was something for the geeks and nerds of the world, not for ordinary people going about their everyday lives. In the space of little more than six months, the world has changed. Anyone with a computer or smartphone can use ChatGPT. Articles in the press have praised and condemned AI in equal measure. In March this year, Elon Musk and other tech giants signed a letter calling for a pause in AI development, while we figure out all the ramifications. But in July – less than six months later – Musk announced the formation of a new company focused on artificial intelligence. It’s hard to keep up!
At SD Law we have been following the growth of generative AI. Generative AI refers to tools that can identify patterns across enormous sets of data and generate new content, something that until now has been considered uniquely human. It’s understandable that we, as lawyers, have a keen interest in AI. A recent report by the McKinsey Global Institute positioned business and legal professionals alongside STEM (science, technology, engineering and maths) professionals as the occupations where generative AI could accelerate automation significantly. These occupations are also classed as resilient and growing. AI won’t put them – or us – out of business. But it will change the way we work. For example, generative AI can help the legal profession with document generation, legal research, contract analysis, due diligence (by analysing large volumes of documents), risk assessment, and other areas. AI can keep us informed about changes in laws by analysing legal databases and summarising relevant changes, and can handle routine client enquiries, freeing up our time for more complex tasks. Clients will benefit from better service and law firms will make better use of valuable resources.
But where are women in all this? AI presents both opportunities and challenges for women and, in any discussion of gender equality, we need to factor in AI and the opportunities and threats it poses to women. There are also some challenges that are unique to the sub-Saharan African context. Let’s explore all the angles.
Reasons for concern
Automation, job displacement and job value
One of the biggest concerns about generative AI, not only for women, is that it will displace jobs through automation. The jobs it will replace are those that contain a high level of repetitive work and require a low level of skill. Therefore, the workers it replaces will not easily transfer to other roles. This concern is more pronounced among women because women make up a higher proportion of the low-skilled, low-paid workforce. Women are overrepresented in sectors like retail and administrative support, which are the most vulnerable to automation.
Of course, there are some activities that can only be performed by a human being, even if they can be assisted by AI. Caring roles, such as nursing, care of the elderly, early childhood development, etc., require hands-on involvement and emotional intelligence, two qualities that can’t be replicated by AI. Even if a chatbot can provide counselling (and the desirability of this is questionable), it can’t change a baby’s nappy or feed a frail old person. Unfortunately, these occupations are already underpaid, arguably because they are dominated by women. There is a risk that the expansion of AI will perpetuate the undervaluation and underpayment of caregiving roles, thus further widening the gender pay gap.
Bias and discrimination
The risk of bias in AI algorithms is not unique to women; we discussed this in AI and ChatGPT – what are the ethical considerations? But there are reasons to be particularly cautious about the use of AI models in the context of gender equality. If not carefully designed, AI algorithms can perpetuate gender biases present in historical data, resulting in discrimination, particularly in recruitment. Amazon tried to build an AI model that would automate its review of CVs and finally gave up because its historical data was disproportionately biased towards male candidates. The software concluded “we don’t hire women” and Amazon was unable to rectify the bias. A less observant (or less scrupulous) organisation might overlook or ignore this bigotry and exclude women from roles for which they are qualified and suitable.
Furthermore, in addition to exclusion, AI models trained on biased data may reinforce stereotypes and discrimination against women, leading to a downward spiral of inequity as managers justify decisions based on “objective” data. Lastly, the underrepresentation of women in AI development and leadership roles may lead to technologies designed around traditional male strengths and priorities and overlook women’s needs and perspectives.
Reasons to be optimistic
We need to be cognisant of these challenges as we move forward with AI – as a society and as employers. However, managed carefully and thoughtfully, AI can provide women with considerable opportunities.
The risk of job displacement through automation is real, but so is the chance to reskill and upskill for more sophisticated roles. While it may not be an easy shift from a traditional administrative task base (for example) into new technology, roles that involve AI development, data analysis, and AI governance will become more widespread and will leverage the transferable skills gained in conventional roles. History has shown that new technologies have not led to unemployment, but rather have simply shifted employment into new occupations. In the long run, technology has been a net creator of jobs. New technology often destroys existing jobs, but it creates new possibilities through different channels.
Closing the gender pay gap and addressing AI bias
While legitimate concerns exist over the widening of the gender pay gap caused by the devaluation of caregiving roles, there is a flip side to the coin. If women reskill and upskill to enter AI-related fields, they can improve their access to high-paying tech jobs, potentially narrowing the gender pay gap. There is considerable scope for women in AI development, which would have broad social benefit. Women’s participation in AI development will help to avoid or eliminate bias and ensure AI systems accommodate diverse needs and experiences. The involvement of women in AI policy and governance formation will ensure that new regulations address gender bias, discrimination, and privacy concerns. The inclusion of more women in AI development and leadership roles will empower other women to harness AI tools for social impact, entrepreneurship, and innovation.
However, women are currently underrepresented in software and AI development and training. Ratios vary in different countries but the gender gap in these fields has been persistent and needs to be proactively addressed. According to various reports, women comprise only 20-25% of the software development workforce in many countries. Women are also underrepresented in AI research and development, accounting for approximately 15-25% of roles, so opportunities abound, but this low representation may present a barrier to women who feel excluded from the tech “boys’ club”.
Uniquely African challenges and solutions
The challenges facing women in AI are more glaring in low- and middle-income countries. AI can amplify existing gender disparities. But it can also be the source of solutions to specifically African problems.
Access to education and training
Women in sub-Saharan African countries often have limited access to quality education and training in AI and related fields. This lack of education can restrict their participation in AI development and use. In South Africa, less than 13% of women choose to study STEM disciplines, compared with 28% of men, according to the Global Gender Gap Report 2021. There are multiple reasons for this, including gender stereotypes that discourage girls from studying maths and technology-related subjects. There is a lack of role models for girl learners from underrepresented and marginalised communities, leading to fewer women pursuing STEM careers. Furthermore, schools need relevant infrastructure, such as science labs and computer centres, and appropriately qualified teachers for STEM subjects. These are very often lacking in disadvantaged communities.
Digital gender divide
Poor access to training in tech subjects goes hand in hand with scarce access to digital technologies and the internet. This limits the capacity for girls to acquire digital skills and engage in AI-related activities. While boys in resource-poor communities experience the same limitations as girls, there is more encouragement for boys to access the opportunities that do exist, though that is changing rapidly, as coding clubs target girls and educators encourage young women to study STEM subjects. However, in rural areas in particular, social barriers influence girls’ and women’s access to AI resources, opportunities, and platforms.
AI in social development
If women in sub-Saharan Africa are encouraged to participate in AI development, they could contribute to solutions that directly serve their communities, particularly in healthcare, education, and agriculture. AI can also create pathways for women in low- to middle-income countries to start tech-based businesses and innovative projects, creating opportunities for entrepreneurship and economic empowerment.
Uses of digital platforms
AI has huge potential for good in the African context, where distances from urban centres and poor transport links can deepen disadvantage. AI-enabled remote work and digital platforms may provide women with flexible income-generating opportunities. AI applications, such as chatbots and mobile apps, can provide women in remote areas with access to vital information, healthcare resources, and educational content. AI tools can help women advocate for their rights, raise awareness of social issues, and connect with global networks.
However, for African women to benefit from AI, the barriers they face must be addressed. This means providing accessible AI education and training, addressing bias in AI systems, supporting women’s entrepreneurship, and encouraging policies that prioritise gender equality in AI initiatives. Collaboration is needed among governments, commercial and non-profit organisations, and the global AI community, in order to harness the positive potential of AI while minimising the potential for negative impact on women.
Time for change
The introduction and uptake of generative AI in the last year proves that social change can happen seismically as well as gradually. Let’s hope, this Women’s Month, that AI can be a catalyst for change for the women of South Africa.
- The legal implications of generative AI in the workplace – part 1
- The legal implications of ChatGPT in the workplace – part 2
- AI and ChatGPT – what are the ethical considerations?
- Women in the workplace: How equal is equal?
- Women’s Rights – How far have we really come?
The information on this website is provided to assist the reader with a general understanding of the law. While we believe the information to be factually accurate, and have taken care in our preparation of these pages, these articles cannot and do not take individual circumstances into account and are not a substitute for personal legal advice. If you have a legal matter that concerns you, please consult a qualified attorney. Simon Dippenaar & Associates takes no responsibility for any action you may take as a result of reading the information contained herein (or the consequences thereof), in the absence of professional legal advice.