Inclusion and Ethics in AI
Artificial Intelligence (AI) is now one of the most important technologies of our time. It is transforming industries and everyday life. Its development, however, has not been perfect, and concerns like ethics and inclusion frequently arise. Africa is replete with innovation and a diversity of cultural perspectives, making these issues even more evident because the continent is often left out of the global AI conversation. What is the Pan-African agenda for addressing the global equity crisis in technology? What are the trust, safety, and accountability gaps? Are there solutions?
The Current State of AI in Africa
AI development in Africa remains a story of untapped potential and systemic challenges. While countries like South Africa, Kenya, and Nigeria lead AI research and startups on the continent, Africa is underrepresented in total global AI activity. For example, Stanford University’s AI Index 2021 Report indicated that only 0.3% of global AI research publications came from Africa. This reflects a staggering underinvestment in intellectual capital on the continent.
In addition, big tech companies have underinvested in trust and safety in markets they deemed smaller, permitting toxic content—including misinformation and manipulative media—to spread unchecked. A prime example was during the Tigray war in Ethiopia. Facebook was criticised for not stopping hate speech and incitement to violence because it did not have content moderators familiar with the local languages. Arguably, the widespread consequences could have been limited if more local content moderators had been utilised. This lends credence to the argument that Africa needs African-specific trust and safety solutions on global platforms.
Bias and Inequality in AI Systems
AI systems are biased because they are trained with data that does not represent everyone. Most AI data are based in the Global North, meaning African realities are often misrepresented or ignored. A study in Nature Communications found that datasets used to train global AI systems are hugely dominated by data from the Global North, with Africa contributing less than 1%.
For instance, facial recognition technology has been shown to perform poorly on people with darker skin. A 2019 study in the United States revealed that these systems were much more likely to misidentify Black and Asian faces compared to white faces. Such biases erode confidence in AI and lead to disparate treatment in areas like law enforcement and identity verification.
To counter such bias, Africa-led initiatives like Masakhane are bridging the gap. Masakhane is a grassroots NLP community tackling the development of AI systems to identify African languages, cultures, and names. Their work includes machine translation of Swahili, Zulu, and Yoruba to ensure African voices and cultures are represented in AI.
Counteracting Harmful Content
The spread of misinformation and synthetic media is a pressing issue in Africa. During the COVID-19 pandemic, misinformation about vaccines was widespread on social media, with disastrous public health consequences. This issue was exacerbated by the failure of technology firms to provide accurate content moderation in local languages.
One response was partnerships like the Africa Check and WhatsApp collaboration. Before Meta announced the end of the partnership in early 2025 (a decision widely criticised), Africa Check worked with WhatsApp to combat misinformation by flagging fake articles on platforms and employing local fact-checkers to provide culturally appropriate, accurate information to African audiences.
However, much more needs to be accomplished. Governments and civil society must come together to institute digital literacy initiatives that empower citizens to differentiate between credible information and disinformation.
The Pan-African Agenda for Ethical AI
The Pan-African agenda for ethical AI is guided by the vision of developing technology infused with African values, supporting equity, and addressing the continent’s unique challenges. Initiatives like the African Union’s Digital Transformation Strategy for Africa (2020-2030) aim to harness digital technologies, including AI, to drive inclusive economic growth.
An example is the establishment of the “Smart Africa” program, which unites 40 African countries in pursuing a single digital communications market. The alliance seeks to foster accountable AI development and gives African nations autonomy to shape their digital policies.
Moreover, countries like Ghana have launched AI research centres, such as the Google Research Centre in Accra and the Ghana Natural Language Processing (NLP) Lab. These centres specialise in integrating local languages into web translators and adapting language technology for local challenges. Such initiatives demonstrate the continent’s active role in developing technology suited to its people.
The Role of Big Tech
Big technology companies have a significant role in ensuring inclusion and ethics in AI. However, their performance in Africa has been less than ideal. In 2021, whistleblower Frances Haugen revealed that Facebook prioritised profits over trust and safety in countries where English was not the primary language, disregarding the need for translations into local languages.
Despite this, there are positive examples. Google’s AI lab in Accra, Ghana, launched in 2019, collaborates with local researchers to develop solutions for African challenges. However, such efforts must go beyond token investments. Tech companies need to engage African stakeholders in decision-making processes to align their operations with local needs and priorities.
Amplifying African Voices in Global AI Discussions
Africa has much to offer in the global discussion on AI ethics. Africa’s diverse cultural perspectives and histories enhance our knowledge of building fair and inclusive AI systems. For example, the Ubuntu philosophy, which emphasises interdependence and communal accountability, can shape ethical AI systems that prioritise societal good above individual interests.
The future of AI in Africa is our collective responsibility. Inclusion and ethics in AI are not just moral imperatives; they are pragmatic imperatives for a technology that aims to serve humankind. By embracing the Pan-African agenda and amplifying African voices, the global AI community can make tangible strides toward a more balanced and accountable future.
