Next week, about sixty of us will gather at AEA Plaza in Nairobi to ask a question that sounds simple and is not: can Africa move from being a consumer of digital technology to a creator of it, and can that shift be guided by ethics, human dignity and the common good?
I have spent the past weeks pulling together the evidence for the opening overview. This post is the short version of what I found, and why I think faith and moral leaders are not spectators in this story but some of its most important actors.
The technology is moving faster than anything before it
Generative AI reached 53% of the world's population within three years of ChatGPT's release, faster than the personal computer or the internet ever spread. Stanford's 2026 AI Index reports that frontier models now meet or exceed human baselines on PhD-level science questions and competition mathematics, and that on one widely used coding benchmark, performance went from 60% to near 100% in a single year. Benchmarks designed to last a decade saturate in months.
The more important shift is from chatbots that answer to agents that act. An agent can be told to reconcile two spreadsheets and email the discrepancies to finance, and it will open the files, write the code and send the mail. Agent accuracy on structured computer tasks rose from roughly 12% to 66% in a year. When software can act, the question changes from "is the answer correct?" to "who authorised this, who is accountable, and what did it do while nobody was watching?" Those are questions the church has a long vocabulary for.
Yet capability is jagged. The same model that won a gold medal at the International Mathematical Olympiad reads an analogue clock correctly only half the time. Documented AI incidents rose from 233 to 362 in a year, and transparency scores for the leading models fell after two years of improvement. Smarter is not the same as more trustworthy.
Africa's bottleneck is not coverage. It is use.
Here is the number that reframed the whole talk for me. Only 9% of Africans live outside mobile broadband coverage. But 63% live under coverage and still do not use the internet. The network is there; the phone, the data bundle, the digital skills and the content in one's own language are not.
And the AI boom is making this worse in real time. The GSMA's report released this week warns that demand for AI data centres has more than doubled memory-chip prices since late 2025, pushing entry-level smartphones out of reach just as they were becoming attainable. For the poorest 20% in Sub-Saharan Africa, a basic smartphone now costs 76% of monthly income. The AI economy is, at this moment, pricing the poorest out of the internet.
The adoption figures follow. Microsoft's diffusion data puts South Africa at 23% of working-age adults using generative AI, Nigeria at 10%, Kenya under 9%, and most of the continent below 8%. The UAE is at 70%. The gap between Global North and South is widening, not closing.
The IMF's July paper is blunt: AI could add about 4% to Sub-Saharan Africa's output over the next decade if electricity, connectivity and skills improve, or a negligible 0.2% if they do not. Half the region still lacks reliable power. "It's hard to have anything without electricity."
But Africa is building, and language is the lever
None of this is a counsel of despair. Cassava Technologies and Nvidia have 3,000 GPUs live in South Africa with 12,000 more planned across Kenya, Nigeria, Egypt and Morocco; Egypt's first national AI factory launched this month. Forty-five African countries now have data-protection laws; sixteen have national AI strategies; Kenya's AI Bill is before Parliament with a proposed AI Commissioner. African startups raised $1.44 billion in the first half of 2026, and AI has become, in one analyst's phrase, "a line item, not a buzzword."
The most hopeful signal is language. Africa is home to more than 30% of the world's languages, and today's models barely speak them. But an African-led ecosystem is filling the gap: African Next Voices building the largest African-language dataset, Lelapa AI's InkubaLM running on cheap hardware, CommonLingua identifying 61 African languages, LINGUA Africa funding applied tools for health workers and farmers. When models improved in Korean, Thai and Japanese, adoption in those countries jumped. The same lever is available to us.
And who on this continent knows more about mother-tongue communication than the church? Bible translation is one of the oldest language-technology enterprises in Africa. The expertise this ecosystem needs already sits in our seminaries and translation offices.
The shadow side is already here
Honesty requires naming two harms that are not hypothetical.
INTERPOL's 2026 assessment finds AI involved in 55% of reported cybercrime in Africa. Losses more than doubled to $484 million. Deepfaked faces are defeating bank biometrics in Kenya and Tanzania. Our congregations are being scammed on the same mobile-money rails that brought them financial inclusion, and when it happens, it is often the pastor they call first.
Then there is the hidden workforce. Kenyan data annotators and content moderators earn between $1.46 and $3.74 an hour to label training data and filter the worst content on the internet, so that AI can be "safe" for everyone else. The same work pays $21 to $27 in the United States. More than a quarter report significant psychological distress. Kenya's draft worker-protection policy is a start, but this is a labour-justice question of exactly the kind the church has spoken to for two centuries.
What the global church is saying, and what is missing
In May, Pope Leo XIV devoted his first encyclical, Magnifica Humanitas, entirely to AI, opening with a choice between a new Tower of Babel and the city where God and humanity dwell together. Lausanne's LIGHT briefing gives evangelicals a balanced posture: AI is "neither savior nor threat in itself; its value depends on how it is discerned, governed, and deployed by God's people." The 2019 Evangelical Statement of Principles anchors everything in the imago Dei.
Across traditions, the common thread is that infinite, non-negotiable human dignity is the fixed point from which every AI question is judged. What is still missing is an African articulation: rooted in Ubuntu and communal life, alert to language, poverty, youth unemployment and extraction, spoken by leaders whom Africans already trust. That, in part, is what the Nairobi forum exists to begin.
Six questions I will put to the room
Dignity. If people are made in God's image, what do we owe the worker paid $2 an hour to view the worst of the internet?
Truth. When a deepfake can defeat a bank and a model agrees with whatever a user believes, how does the church remain a community of truth-telling?
Community. If AI companions deepen loneliness for the most isolated, as a Stanford study found last month, what does incarnational presence now mean?
Justice. Who bears the cost of the transition: the graduate whose entry-level job is automated, the grandmother scammed via mobile money, the language never modelled?
Voice. The Gospel came to Africa in the mother tongue. Will African languages and moral traditions shape the systems our children will learn from?
Stewardship. Sermons, counselling, Bible study: what should pastors delegate to a machine, and what must remain the labour of a shepherd who knows the sheep? I explored this last question at length in Does agentic AI introduce a new kind of mediation into Christian practice?
The scarce resource is trust
If I had to compress the whole overview into one claim, it would be this. In an era where capability is becoming a commodity produced by half a dozen companies, the scarce resource is trust, and African faith communities hold more of it than any lab. The technology will arrive either way. Whether it arrives as Babel or as blessing depends on who is in the room when it is built.
I hope to see you in the room. The full overview, with sources and speaker notes, is at acalytica.com/africa-ai-faith-2026.
