Responsible AI: Gender Inclusion in Africa
By Sylvia Mukasa*
Source: ipopba, Getty Images
Responsible Artificial Intelligence (AI) development and deployment in Africa must address existing AI ecosystem challenges, which include: lack of AI experts/lecturers; limited number of educational institutions; poor funding for AI research, infrastructure and entrepreneurship; exclusion of marginalised populations; algorithmic bias; and digital colonialism. This article focuses on the challenge of gender inclusion in a predominantly male-dominated AI community.
Defining responsible AI
There are many definitions of Responsible AI. In simple terms, Responsible AI can be defined as a governance framework that documents how an organisation is addressing the challenges around AI from both ethical and legal points of view. Accenture defines responsible AI as the practice of designing, developing, and deploying AI with good intention to empower employees and businesses, and fairly impact customers and society.
These notions of ethical and accountable AI have been adopted by many stakeholders from government, industry, civil society and academia. Making AI systems transparent, fair, secure and inclusive are core elements of widely asserted responsible AI frameworks.
Calls for the development of responsible governance approaches to harness AI technologies responsibly are already underway. This may include collaborations to support civil society organisations, digital rights labs and African innovators in their efforts to build governance accountability models that meet the ethical needs of African democracies.
Understanding AI and gender inclusion in Africa is important to help mitigate any gender-related exclusion. We will focus on skilling, entrepreneurship, human rights and governance frameworks as key drivers to inclusive AI and how this affects/applies to Africa.
Technology entrepreneurship, skilling and gender
Despite the recognition of a skills gap in Africa, gender disaggregated data on the actual state of AI skills is scanty. There is over-dependence on proxy indicators like the Government AI Readiness Index due to few tools to measure capacity and limited data availability on Africa. Use of proxies is problematic since this may not give disaggregated details to give a clearer picture of the situation. The extent to which and how AI technology is being utilised within programmes aimed at boosting employment or building workforce capacity by gender is equally unclear.
Gender inequality remains an entrenched reality for women in many African settings; however, African entrepreneurship is experiencing a transformative “feminisation” of technology entrepreneurship. According to a Mastercard Report, the gender gap is disappearing in some countries like Angola, Botswana, Nigeria, South Africa and Zambia.
These countries and others, like Kenya and Rwanda, are adopting proactive policies to increase the number of girls taking up STEM subjects and computer science, with support from technology companies. African women are increasingly engaging with code through initiatives like South Africa’s Code4CapeTown. SingularityNET, an AI-focused start-up, is at the forefront of hiring and promoting African female engineers and Kenya’s GlobalX Innovation Labs is pushing for inclusive AI through women-focused initiatives.
This is not to say that the gaps in tech entrepreneurship and skills are resolved. Far from it. Element AI research shows that globally, only 12% of leading machine learning researchers; and only 18% of authors at leading AI conferences are women. The AI start-up scene is not spared from female under-representation.
AI start-ups and gender
Emerging AI start-up ecosystems are supporting Africa’s women. An example is the Africa AI Accelerator (AAIA).
AAIA supports the scaling of the next generation of Africa’s AI start-ups. The infographic below shows the data of 10 participating start-ups from four countries: Ghana, Uganda, South Africa, and Rwanda. From their websites, of the 10 selected start-ups, the majority (68%) of the founding partners are male (15 identified compared to 7 female co/founders). Overall, there were twice as many males (12) as females (6) participating in the accelerator’s bootcamp. Another infographic of the programme showed that 33% (17 out of 34) of the applicants had a female founder.
Source: Africa AI Accelerator Facebook
Human rights and algorithmic bias
Gender inclusion advocates are helping to address issues around how data and algorithms disfavour women. For instance, the work of Joy Buolamwini, Timnit Gebru, and Deb Raji, three computer scientists whose work revealed that facial recognition systems were far less accurate on Black female faces than white male ones. AI-powered recruitment software has been found to discriminate against women. Due to limited research, the extent to which these flaws affect Africans remains largely unknown.
Governance frameworks
Over 50 African countries still lack national AI strategies. This means limited preparation of appropriate regulatory and ethical frameworks. Yet, without these, gender inclusion cannot be achieved. As countries continue on their journey to develop these frameworks, a key question arises: What can be done to advance gender inclusion?
Ways forward for Africa
African ecosystem players like the Africa Union’s AI working Group can work in collaboration with international organisations like the United Nation to develop AI strategies and frameworks, identify regulatory and governance issues and learn best practices.
The United Nations Office of the Human Rights High Commissioner provides recommendations on how human rights considerations can be backed into the development and deployment of AI. These include the development of human rights impact assessments on AI technologies and the need to adopt responsible policies to prevent the violation of 85 human rights.
UNESCO’s Global Dialogue on Gender Equality and AI with leaders in AI, digital technology and gender equality from academia, civil society and the private sector pointed out the need to:
Generate an understanding of AI Ethics Principles and how to position gender equality within them.
Reflect on possible approaches for operationalising AI and Gender Equality Principles.
Applying a gender lens whilst investing in technology entrepreneurship and establishing accelerators/incubators, equipping women with technical skills and the knowledge and training to govern the deployment of AI is paramount.
Job automation through AI risks having a negative impact on women’s economic empowerment and labour market opportunities. Equipping women with future skills is therefore vital. On a positive note, AI can also be a solution for advancing gender equality. For example, AI-powered gender-decoders help employers use gender-sensitive language to write job postings that are more inclusive in order to increase the diversity of their workforce.
The use of unbiased AI solutions is the only way to deal with algorithmic bias. There is need for Africa-focused gendered research to determine the extent to which Africa’s women are affected by algorithmic bias, for example in terms of hiring and getting online gigs. Increasing AI-gender-related advocacy work by Africans and for Africa to get the attention of AI ecosystem players on the gaps and challenges will definitely boost inclusive AI on the continent.
In conclusion, unless AI technologies are developed and deployed in an equitable and ethical manner, there is reason to believe that they will further entrench the gender divide in Africa and globally. This divide could be seen in terms of the number of women in AI, either as professionals or entrepreneurs. It could also mean a widening AI skills gap, less consumption and building of AI solutions by women and; no or inadequate gender-equality policies, ethics and frameworks to ensure protection and inclusion of women in matters AI. We need to ensure that no-one is left behind in the digital era.
*Sylvia Mukasa is an award-winning entrepreneur. Sylvia is Founder/CEO of GlobalX Investments Ltd/GlobalX Innovation Labs which focuses on emerging technologies. She is passionate about empowering Women in Tech and contributing to the entrepreneurial ecosystem in Africa and globally. She is Country Co-Founder/Chapter Lead (Kenya) for Women in Tech Africa (WiTA). WiTA won the United Nations EQUALS in Tech Award, Leadership Category in 2018. She is a 2014 TechWomen Fellow, an Initiative of the U.S. Department of State's Bureau of Educational and Cultural Affairs. TechWomen empowers, connects and supports the next generation of women leaders in STEM from Africa, Central and South Asia, and the Middle East; launched by Hillary Clinton, former US Secretary of State. Sylvia holds a certificate from the GIZ/ITCILO’s Leading with AI Lab. She is a member of the Gender Alliance, the Global Leadership Academy (GLAC) and a BMW Foundation Herbert Quandt Responsible Leader.