Artificial intelligence is often discussed as if its main challenges were technical: better models, faster systems, more accurate outputs. But every system also reflects assumptions about what counts as a good decision, whose interests matter and how individual choices relate to wider communities.
Ubuntu Changes the Starting Question
In an essay for Aeon, philosopher Fainos Mangena argues that AI developed for African societies should take Hunhu/Ubuntu ethics seriously. The point is not simply to give existing systems an African surface. It is to ask whether the values built into those systems should be different from the start.
Ubuntu is often associated with a relational understanding of personhood: individuals exist within networks of obligation, care and community. That creates a different ethical emphasis from models that place individual autonomy above nearly everything else.
AI Is Never Culturally Neutral
Software can appear neutral because its decisions are expressed through numbers, interfaces and automated outputs. Yet choices enter at every stage: what data is collected, what behaviour is rewarded, what risks are tolerated and whose definition of fairness is used.
This is where the debate connects with public philosophy and everyday questions. A recommendation system, hiring tool or educational model is not only solving a technical problem. It is making assumptions about people.
Local Values Need More Than Translation
There is a temptation to treat localisation as a language problem: translate the interface, add regional data and the job is done. Mangena’s argument points to something deeper. A system can speak a local language while still carrying imported ideas about responsibility, success and social relationships.
That does not mean there is one single African ethical model, nor that Western and African traditions fit into simple opposites. The more useful question is whether AI development leaves enough room for different societies to contest the values embedded in the technology they use.
The debate over an “AI for Africa” is therefore larger than geography. It asks who gets to define what intelligent systems are for.
