Imagine an African city of three million people preparing to deploy an artificial intelligence (AI) system to manage traffic and public safety.
Sounds irresistible: cleaner transport, lower emissions, faster emergency response, and public services that anticipate problems rather than waiting for them to happen.
Then comes the other side of the deal.
The system depends on widespread facial recognition, predictive policing and algorithms that analyse citizens' data to identify possible crime patterns.
The city may become more efficient. It may also become a place where residents are constantly being watched, classified and assessed by systems they may not fully understand.
That dilemma came into sharp focus last week, September 21, at Digital@UNGA 2026, during a session titled Human-Centric Cities in the Age of AI: Connectivity, Inclusion and Opportunity for All.
The discussion examined how AI can transform cities while keeping people, rather than technology, at the centre.
The fictional city presented during the session captured a dilemma that is becoming increasingly real across Africa.
From a private-sector perspective, the attraction is obvious: build a system that improves transport, public safety and environmental outcomes while creating a viable business.
But the same person selling the technology is also a citizen.
Their children live in the city. Their grandchildren may live there.
What happens when their faces, fingerprints and other personal information become part of a database whose future uses, security and access they cannot fully control?
That is the harder question in Africa's AI conversation.
Africa does not only need AI. It needs African AI
The question comes at a time when Africa is being encouraged to build its own AI ecosystem rather than remain a passive consumer of technology developed elsewhere.
The United Nations Economic Commission for Africa (UNECA), working with the International Telecommunication Union (ITU) and Research ICT Africa, is inviting African innovators to showcase functioning AI-enabled solutions at the inaugural ECA Innovation Cafe during the Africa Internet Governance Forum in Accra, from November 2 to November 5 2026.
The call is open to startups, universities, research institutions, innovation hubs, public and private institutions, social enterprises and individual innovators developing solutions to African development challenges.
Selected innovators can demonstrate their products to policymakers, investors and other stakeholders, build partnerships, and explore opportunities for wider adoption.
The initiative is more than technology. It asks what African innovators can build for African problems.
That opens a more interesting set of questions. Could an AI system better understand county budgets? Could it make parliamentary records searchable in English and Kiswahili?
Could it help journalists search decades of radio archives? Could it assist teachers with locally relevant learning material? Could it help farmers interpret weather information?
Could health workers use AI as an additional tool without surrendering professional judgement?
The opportunity may lie in making AI understand Africa better. But that requires data.
And data requires trust. The citizen cannot be an afterthought.
Kenya provides an important example
The country's AI Strategy 2025–2030 identifies AI infrastructure, data and AI governance, research, innovation and commercialisation among its core areas. The Government is also developing a national policy for artificial intelligence and other emerging technologies, with the draft framework addressing responsible development and deployment, innovation, infrastructure, human capital and public trust.
For Data Commissioner Immaculate Kassait, the AI conversation has never been simply about clever machines.
It is also about the information those machines consume.
At the Kenya AI and Software Summit in 2025, Kassait emphasised transparency in how data is collected, processed and safeguarded, describing trust as essential to digital transformation.
That principle becomes particularly important when AI moves from a chatbot on a person's phone into the infrastructure of a city.
A facial-recognition system can help identify a person. A predictive system can flag a pattern. An automated system can process information at a scale no human team could easily match.
But who checks it? Who can challenge its decision? Who owns the underlying data?
How long is it retained? What happens if it is breached?
These are not merely technology questions. They are questions of governance.
Kenya's Constitution protects privacy under Article 31, while the Data Protection Act, 2019 provides the statutory framework for the processing and safeguarding of personal data.
The emerging AI policy framework adds another layer to that governance architecture.
The challenge is ensuring that regulation does not arrive only after technology has already become embedded in people's lives.
Then there are children
The human-centred test becomes even more important when the people affected are children.
In January 2026, the ITU, UNICEF, UNESCO, the UN Committee on the Rights of the Child and other international organisations launched a joint statement calling for a child-rights-based approach to artificial intelligence.
The statement says AI should be designed, developed, deployed and governed in ways that respect, protect and promote children's rights, highlighting privacy, safety, non-discrimination, inclusion, transparency and accountability.
UNICEF's guidance similarly calls for child-centred AI that protects children's data and privacy, prevents discrimination and provides transparency and accountability.
Africa has also been developing its own response.
The African Union's Child Online Safety and Empowerment Policy, endorsed in 2024, recognises threats to children's privacy, identity and agency arising from personal data, biometrics and automated decision-making. It calls for child-centred design and stronger responsibility from companies alongside legal, regulatory and educational measures.
The debate is increasingly moving towards a simple principle: technology providers should bear responsibility for making products safer before children encounter harm, rather than leaving families and children to deal with the consequences afterwards.
That principle matters beyond social media.
It applies whenever AI systems determine what children see, how they are profiled or how their data is used.
Kenya's election makes the question harder
As Kenya moves towards the 2027 General Election, AI is also becoming part of political communication.
The ODPC is currently seeking public comments on a draft notice proposing political affiliation as an additional category of sensitive personal data in specified cross-border-transfer circumstances.
The notice says political affiliation includes information from which political leanings can reasonably be inferred. The deadline for comments is September 25, 2026.
AI systems can process enormous quantities of information. When that information includes political preferences, or signals from which political leanings can be inferred, technology can potentially change the scale and sophistication of political targeting.
A recent example came from Anthropic, the company behind Claude, which said it had identified and disrupted a Kenya-based operation that used its AI system to generate political content at scale.
The case, as reported by The Standard, included material praising Energy CS Opiyo Wandayi and content portraying opposition figures as politically weakened.
It is one example, not proof that AI has transformed Kenya's entire political environment.
But it demonstrates why data protection, transparency and accountability matter as the country approaches another election.
The newsroom is a smaller version of the same problem
This is another place where the human-centred AI debate is already playing out.
AI can transcribe interviews, search lengthy reports, organise documents, identify patterns in datasets, translate material and suggest headlines.
For a small newsroom working under tight deadlines, that can be valuable.
But the Media Council of Kenya (MCK) has placed boundaries around its use.
Its 2025 Code of Conduct for Media Practice requires AI-generated or technology-assisted content to undergo human editorial review before publication. It also addresses accuracy, intellectual property, harmful stereotypes and disclosure where AI has been used to create or modify editorial images, video or other content.
The principle is straightforward.
Use the machine to process information. Do not outsource responsibility for the journalism.
The same principle applies beyond journalism.
An AI system can help a city manage traffic. It can help a hospital organise information. It can help a government identify patterns. It can help a teacher prepare material.
But the fact that a system can make a decision does not automatically mean it should be allowed to make that decision without meaningful human oversight.
This is why initiatives such as the ECA Innovation Cafe matter.
Africa needs innovators who understand its languages, infrastructure gaps, markets and communities. But African innovation should not be measured only by whether a product works.
It should also be tested against harder questions. Who benefits? Who could be excluded? Whose data is being used? Can people understand how the system affects them? Can they challenge a decision? What happens to children? What happens to people poorly represented in the data?
And what happens when a system designed to solve one problem creates another?
The answers will not come from technology alone.
They require governments, regulators, companies, researchers, journalists and citizens to participate in the conversation.
At the UN, Secretary-General António Guterres has called for global cooperation on AI safety and stressed the importance of systems that are safe, transparent and accountable and keep human dignity at their centre.
That is also the message emerging from the human-centric cities discussion.
Connectivity matters. Innovation matters. Opportunity matters.
But so does the person sitting behind the screen, walking through the city, sending a child to school, visiting a hospital, casting a vote or simply trying to live a private life.
Africa does not have to choose between innovation and rights.
The harder task is to build systems in which the two develop together.
The machines will become more capable. African cities will become more digital. More businesses and public institutions will use AI.
The question is no longer whether Africa will enter the AI age. It is what kind of AI age Africa will build.
The strongest test may not be whether the machine can see more, process or predict more.
It may be whether, in the rush to build smarter cities and smarter systems, Africa remembers the people those systems were supposed to serve. Something to ponder.
amurumba@standardmedia.co.ke
