Blab AI
Generative AI and Ubuntu: A Study of Relational Apartheid in South Africa
When a chatbot says ‘I’m sorry’, does it feel the weight of that apology? Generative artificial intelligence has slipped into everyday life, powering assistants that tutor, serve, and even offer a virtual shoulder. A recent academic study examines how these AI‑mediated interactions intersect with South Africa’s legacy of apartheid and the African philosophy of ubuntu.
Ubuntu frames personhood as a web of relationships: recognition, care, responsibility, and shared life. The research argues that the rise of generative AI is not merely a technical evolution but a relational one. It introduces the term relational apartheid to describe a pattern where meaningful human engagement is unevenly distributed. Some users are greeted by people; others are managed by systems.
Large language models can generate fluent, emotionally sensitive responses—apologising, encouraging, consoling—remembering details within a conversation and adjusting tone. For many, these exchanges feel surprisingly human. Yet the study stresses that a chatbot cannot share in a person’s life, cannot be vulnerable, cannot be held accountable, and cannot experience moral burden or transformation. Human relationships involve mutual risk, disappointment, misunderstanding, repair, and growth—elements that AI lacks.
The concern is that AI’s convenience may train users to expect relationships without the hard work of reciprocity. While AI tools can help people find information, write better, learn faster, and access services, the danger arises when AI substitutes for human presence in areas where recognition matters. Customer service offers a clear example: as organisations automate front‑line support, callers increasingly encounter chatbots before reaching a human. This creates a tiered system in which some customers receive human attention while others are left with automated interaction.
The research cites Salesforce, a major provider of customer‑service software, reporting that AI agents now handle a growing share of customer interactions and that the company has reduced thousands of support roles in recent years. The study cautions that while not all customer‑service work will disappear, routine service tasks can be reorganised quickly once AI becomes the default front line.
Similar dynamics could unfold in healthcare, education, and social support. Where human professionals are scarce, AI counselling tools, tutoring systems, and advice bots may appear to offer practical solutions. They can help, but they also risk normalising a situation in which underserved populations are increasingly spoken to by machines rather than people. In a society described by the World Bank as among the most unequal in the world, old differences in income, education, language, geography, and institutional power could re‑emerge in new digital forms.
The study also examines efforts to embed ubuntu into AI design. Scholars argue that designing AI systems that are more sensitive to African languages, local histories, and communal values, and involving communities in decisions about how AI is built, can add value. However, the research points out that ubuntu is not merely a set of polite phrases or cultural preferences; it is a way of thinking about persons in a relationship that depends on shared life, mutual vulnerability, and accountability. Current AI systems cannot live these values in the way people do.
Policy implications follow from this distinction. AI systems should be presented clearly as tools, not companions. They should not blur the difference between simulated care and real care, especially when used by children, the elderly, patients, or socially isolated users. In sensitive settings, AI should support human professionals rather than replace them.
The study concludes that generative AI may help South Africa improve public services, widen access to knowledge, and support overburdened institutions. Yet it may also deepen the distance between people if efficiency becomes the main measure of progress. The research calls for careful governance that recognises the relational nature of human dignity and guards against a new form of digital inequality.
Today, generative AI tools are widely deployed across customer‑service, healthcare, and education sectors in South Africa. The study highlights the need for policy frameworks that treat AI as a supportive tool, preserve human‑to‑human interaction where dignity is at stake, and ensure that the benefits of AI do not reinforce existing social divides. The research remains a call to embed ubuntu‑aligned principles into AI design and deployment, acknowledging that while technology can simulate empathy, it cannot replace the relational depth that underpins human dignity.
Ubuntu frames personhood as a web of relationships: recognition, care, responsibility, and shared life. The research argues that the rise of generative AI is not merely a technical evolution but a relational one. It introduces the term relational apartheid to describe a pattern where meaningful human engagement is unevenly distributed. Some users are greeted by people; others are managed by systems.
Large language models can generate fluent, emotionally sensitive responses—apologising, encouraging, consoling—remembering details within a conversation and adjusting tone. For many, these exchanges feel surprisingly human. Yet the study stresses that a chatbot cannot share in a person’s life, cannot be vulnerable, cannot be held accountable, and cannot experience moral burden or transformation. Human relationships involve mutual risk, disappointment, misunderstanding, repair, and growth—elements that AI lacks.
The concern is that AI’s convenience may train users to expect relationships without the hard work of reciprocity. While AI tools can help people find information, write better, learn faster, and access services, the danger arises when AI substitutes for human presence in areas where recognition matters. Customer service offers a clear example: as organisations automate front‑line support, callers increasingly encounter chatbots before reaching a human. This creates a tiered system in which some customers receive human attention while others are left with automated interaction.
The research cites Salesforce, a major provider of customer‑service software, reporting that AI agents now handle a growing share of customer interactions and that the company has reduced thousands of support roles in recent years. The study cautions that while not all customer‑service work will disappear, routine service tasks can be reorganised quickly once AI becomes the default front line.
Similar dynamics could unfold in healthcare, education, and social support. Where human professionals are scarce, AI counselling tools, tutoring systems, and advice bots may appear to offer practical solutions. They can help, but they also risk normalising a situation in which underserved populations are increasingly spoken to by machines rather than people. In a society described by the World Bank as among the most unequal in the world, old differences in income, education, language, geography, and institutional power could re‑emerge in new digital forms.
The study also examines efforts to embed ubuntu into AI design. Scholars argue that designing AI systems that are more sensitive to African languages, local histories, and communal values, and involving communities in decisions about how AI is built, can add value. However, the research points out that ubuntu is not merely a set of polite phrases or cultural preferences; it is a way of thinking about persons in a relationship that depends on shared life, mutual vulnerability, and accountability. Current AI systems cannot live these values in the way people do.
Policy implications follow from this distinction. AI systems should be presented clearly as tools, not companions. They should not blur the difference between simulated care and real care, especially when used by children, the elderly, patients, or socially isolated users. In sensitive settings, AI should support human professionals rather than replace them.
The study concludes that generative AI may help South Africa improve public services, widen access to knowledge, and support overburdened institutions. Yet it may also deepen the distance between people if efficiency becomes the main measure of progress. The research calls for careful governance that recognises the relational nature of human dignity and guards against a new form of digital inequality.
Today, generative AI tools are widely deployed across customer‑service, healthcare, and education sectors in South Africa. The study highlights the need for policy frameworks that treat AI as a supportive tool, preserve human‑to‑human interaction where dignity is at stake, and ensure that the benefits of AI do not reinforce existing social divides. The research remains a call to embed ubuntu‑aligned principles into AI design and deployment, acknowledging that while technology can simulate empathy, it cannot replace the relational depth that underpins human dignity.
