Deliberate, inclusive AI policies to empower women in Africa 

One year ago, on International Women’s Day 2023, the OECD.AI team asked whether rapidly advancing AI systems addressed longstanding gender equality issues. In the quest for gender equality, the emergence of artificial intelligence (AI) offers unprecedented opportunities and challenges. As nations around the globe celebrate International Women’s Day 2024, AI’s potential to revolutionise society’s fabric is a central theme. Yet, this technological evolution also poses the risk of deepening existing inequalities. This makes examining how AI can catalyse empowering women, with a particular focus on the African continent, critical. 

This year, in honour of International Women’s Day, we share some highlights from a panel discussion on how AI can empower women in Africa as part of the OECD-African Union (AU) AI Governance Dialogue.  

A catalyst for international dialogue on AI governance 

On 6 March 2024, the OECD-AU AI Governance Dialogue, supported by the UK, highlighted the need for truly global cooperation to foster trustworthy AI. The dialogue underscored the essential question of how policies can foster the use of AI to empower women while mitigating the technology’s risks. Such international dialogues are critical in shaping inclusive AI Futures and ensuring AI is trusted across diverse local contexts.  

African voices of change and challenge 

Prominent figures offered insights into the complex interplay between AI and women’s empowerment in Africa, highlighting obstacles and opportunities. 

The discussion’s moderator, Nagham El Houssamy, Associate Director for Research in Middle East and Africa at the Access to Knowledge for Development Center at the American University in Cairo‘s School of Business, questioned whether current AI technologies are addressing gender equality issues or exacerbating them, emphasising the necessity for policy interventions to ensure AI becomes a tool for empowerment rather than exclusion. 

“It is no secret that AI tools are transforming the world in significant ways. Amid this transformation, we must ask: Is today’s AI addressing the gender equality issues that have plagued policymakers for decades? Is it making things worse? How is this playing out in practice in Africa? Are policymakers across the continent doing enough to ensure that today’s AI systems do not perpetuate yesterday’s harmful biases?  

In OECD countries, more than twice as many young men than women aged 16-24 can programme, an essential skill for AI development. Across Namibia, Nigeria and South Africa, women contribute to only about one-third of AI scientific publications. In South Africa specifically, men are twice as likely as women to report AI-related skills on LinkedIn. The fact that the gender gap persists when we look at AI is unsurprising, and we know there is more work to do.” 

Rachel Adams of the African Observatory on Responsible AI underscored the emerging initiatives across the continent to bridge the gender gap in AI. She pointed out the greater social benefits of involving more women in AI innovation and policymaking, stating, “Technology should be socially beneficial…and history shows that the net social benefits tend to be greater if women are centrally involved.” 

Without question, we need more women at the forefront of AI innovation and policy. Technology should be socially beneficial, and the history of peacekeeping and development shows that the net social benefits tend to be greater if women are centrally involved in leadership positions. We need to change the incentives driving the field and industry of AI, which are leading to technology whose benefits are unequally distributed.  Female AI developers often approach AI solution-making because of something they have personally have, not because they see a market opportunity. For example, in Uganda, AI-enabled drones deliver HIV medicine to reach people on isolated islands, chatbots in South Africa help women who were victims of domestic violence navigate the legal system, and AI interfaces can help women access sexual advice outside of traditional institutions that often carry stigma. All of these are examples of why women in Africa need to be a central part of AI discussions.  

This AU-OECD event is a great initiative because we face the same set of questions: how to make AI useful to people and how to scale it responsibly while evaluating its impact on governance and policy. This engagement should not be a one-off event but should be the starting point of structured future conversations where knowledge and information are shared” 

Conrad Tucker of Carnegie Mellon University, Africa (CMU Africa), advocated for increasing women’s representation in AI education and leadership, highlighting the economic and social benefits of engaging the entire population in the AI transformation. 

“About 27% of our student population at CMU Africa are women, but we can do much better than that. We need strategies as an academic institution to expand access for our female student population. When we think about AI and capacity building, we also need to think about employment opportunities for women in AI after they graduate and building the ecosystem for their future success. We need more young female leaders, having a seat at the table is crucial for them as we promote the next generation of female AI leaders, a task that requires to be proactive and intentional.  

At the end of the day, we want our communities to succeed, and it’s simply smart and makes economic sense to not leave out half of the population from the coming AI transformation. Men need to listen more; we do not have the answers related to the female perspective on AI in Africa. Finally, we should not underestimate the power of culture. In Africa there are many cultural aspects that still challenge the advancement of women in the AI field, and as a culture we need to move beyond this.” 

Both Linda Bonyo, Founder of The Lawyers’ Hub, and Peter Addo of Agence Française de Développement drew attention to the systemic barriers hindering women’s participation in AI, calling for a comprehensive approach that tackles these challenges locally and globally. 

Linda Bonyo stated: “One of the hardest things to do for female African AI founders is raising capital. The UN interim report on the state of AI validated this: there is a large capital and talent divide between the Global North and the Global South. In Africa, we also often see a significant gender pay gap, for instance when you look at male and female software engineers who work remotely for the same Silicon Valley company. 

Ghana has the highest number of female entrepreneurs, but the least amount of capital on the African continent. Of course, the AU has made strides in legislation related to gender, such as the Maputo Protocol, but it is often not reflected on the national level. Women in politics face cyber bullying, deep fakes, and political misinformation, often powered by AI, which shows the need for innovative policy frameworks that can keep up with the pace of technology. We also need policy tools for more representation of female voices in global AI conferences and discussions, people need to move at the speed of technology, and we can learn from European initiatives such as Horizon Europe for research or Erasmus for education. Female participation needs to be improved across the board, this includes such simple things as travel support and visas.  

For a long time, data has been manipulated for political purposes, in AI these inequalities are exposed. At the same time, we should not attempt to ban everything but rather look at policy as an incentive for AI development. We have a different context than Europe where discussions about data protection are over 50 years old. We have our own context. What the OECD could do is to use the frameworks on AI policy it already has and bring more African partners and women to join its expert groups to speak about the unique challenges women face on the African continent.” 

Peter Addo also shared perspectives on AI for women’s empowerment in the development context:  

“At the Agence française de développement (AFD), we see a lot of persistent issues for women on the African continent and we need to deplore AI to address these for the benefit of women. For example, over 89% of people working in the informal economic sectors are women, which leads to many problems including access to financial instruments and capital, the provision of health care and insurance, issues of land, or inheritance. There is also the problem of gender-based violence. 

For all these problems, we need to leverage AI and develop inclusive AI solutions that enable women to overcome the hurdles that still prevent them from realising their full potential. This includes mentorships and training women in data science and the fundamentals of AI, not just at the university level, but earlier at high-school and primary school levels.  

The OECD.AI Observatory is a great place for students and teachers to access data and learn about best practices, which can then be adapted to the local context. Another important issue is data and especially bias in data that can fundamentally disadvantage women when developing AI solutions. For this, we need responsible data stewardship at all levels, and data collaboratives between statistical institutions and private companies to leverage this data for the public good.” 

A unified call to action for more women’s representation in African AI discussions 

The panel discussion concluded with an intervention from Christine Nyaguthii, an AI developer from Ethiopia who is pursuing her Master’s in AI Engineering at CMU Africa. She pointed out that guardrails and bias prevention should be a central component of AI development, not an afterthought. She also highlighted the need for proactive policy and the incorporation of young female AI leaders, engineers, and lawyers before a deployed AI system can disadvantage women. 

Conrad Tucker summarised the spirit of the panel discussion well: “We need more Christines and to provide them with platforms such as this AU-OECD panel to act as a force multiplier so female AI leaders in Africa can have an impact on the continent”. 

Looking forward, the call to action is clear: Harnessing AI in the service of women’s empowerment in Africa demands deliberate, inclusive policies and a concerted effort to dismantle the barriers to women’s participation and leadership across technology ecosystems. This young female AI developer’s engagement and emphasis on the importance of preventing bias and including female perspectives in AI development highlight the urgency of this call. 

The OECD AI Policy Observatory as a forum for global AI governance 

The OECD-AU AI Governance Dialogue facilitated what many viewed as just the beginning of the conversation. With clear interest from the OECD.AI community to continue learning about how Africa is tackling the key AI questions of our time, including harnessing AI for women’s empowerment, participants looked forward to the next steps in collaboration and partnerships with the African Union and countries across the continent. 

The OECD Secretariat thanks the governments of the United Kingdom and the African Union Commission for supporting and facilitating this dialogue.