AI: Revolution or hype? Separating fact from fear

NAIROBI, Kenya, Sep 1 — Artificial intelligence has arrived in Kenya long before the country has fully figured out what to do with it.

From banks and insurance companies to technology firms, newsrooms and small businesses, Kenyans are increasingly turning to AI tools to write, translate, analyse information, generate images, assist with coding and automate routine work.

The scale of adoption is striking. The Digital 2026 Mid-Year Global Update, based on research covering more than 240,000 people across 54 markets, found that 97.5 per cent of Kenyan internet users aged 16 and above had used at least one AI tool in the previous month — the highest rate among the markets tracked. Kenya was followed by the United Arab Emirates at 94.2 per cent and Indonesia at 93.6 per cent.

But beneath the excitement lies a more difficult question: Is artificial intelligence delivering the economic revolution being promised, or is much of the current enthusiasm simply hype?

For businesses, AI promises lower operating costs, greater productivity and new products. For workers, it has revived fears that machines could make some jobs obsolete. For consumers, it has opened new concerns around misinformation, privacy and accountability when automated systems get things wrong.

Kenya is therefore moving into a more consequential phase of the AI debate — one that is no longer about whether the technology matters, but whether widespread adoption can translate into measurable economic gains.

From experimentation to impact

Kenya’s enthusiasm for AI is difficult to dispute.

Technology Service Providers of Kenya (TESPOK) Chairman James Turuthi says Kenyans are adopting AI at extraordinary speed, but cautions against confusing widespread use with economic transformation.

“Kenyans are using AI faster than almost anyone on earth. That part is real. What’s still catching up is whether that usage has actually changed how businesses make money,” Mr Turuthi says.

Inside businesses, the picture is more complicated.

Many companies remain at the pilot stage, testing AI for customer service, marketing, fraud detection, credit scoring, document processing and data analysis.

The real measure of success, however, is not whether an organisation has introduced an AI tool. It is whether that tool saves money, generates revenue, improves decision-making, increases productivity or delivers a better service.

PwC’s 2026 Kenya CEO Survey points to a similar challenge, identifying a significant gap between AI readiness and execution and highlighting the need for greater investment in AI skills and innovation ecosystems.

The wider business picture is equally cautious. PwC’s global survey found that chief executives are investing in AI even as immediate returns often remain elusive.

Turuthi says fewer than a quarter of chief executives who have invested in AI report increased revenue, with a similar proportion reporting meaningful cost savings.

“The gap between ‘we’re using AI’ and ‘AI moved our bottom line’ is where I’d tell people to keep their expectations honest,” he says.

There is another problem: Kenya is largely an AI consumer rather than a producer.

Turuthi estimates that only about one in 10 Kenyan organisations are developing AI capability in-house, with most relying on vendors, cloud platforms or AI features already embedded in software they use.

“That’s a sensible, low-cost way to get started. But it also means the ‘Kenya as an AI powerhouse’ narrative is ahead of where we actually sit in the value chain,” he says.

The distinction matters because widespread use does not automatically translate into economic value.

A country that primarily consumes AI products can benefit from improved productivity, but much of the intellectual property, technology ownership and financial value may continue to accrue elsewhere.

What AI can — and cannot — do

The appeal of generative AI lies partly in its ability to perform tasks that previously consumed hours of human labour.

It can draft correspondence, summarise lengthy documents, produce marketing material, translate text, transcribe meetings, analyse information and assist programmers.

For a small business owner, AI could reduce the need to outsource simple promotional material. For an accountant or lawyer, it can reduce the time spent sorting through documents. For a newsroom, it can assist with transcription, translation and research.

But the technology has significant limitations.

AI can produce information that sounds convincing but is false. It can misunderstand context, reproduce biases contained in its training data and provide confident answers even when it lacks sufficient information.

Turuthi argues that AI is most valuable when used for specific, repetitive tasks rather than treated as an all-knowing machine.

Among the practical applications he cites are fraud detection in mobile-money transactions, customer-service chatbots, agricultural tools that identify crop diseases from photographs and systems that assist with transcription and translation between English, Kiswahili and other languages.

“These are real, tested uses, not hype. Use AI as a smart assistant, not an oracle. Verify its output, especially for anything involving money, health, or legal decisions,” he says.

That distinction is increasingly important as AI moves into sensitive areas of everyday life.

Will AI take Kenyan jobs?

Few questions generate as much anxiety as the possibility of mass job losses.

AI can already perform portions of jobs that traditionally required human labour, including data entry, routine customer queries, basic translation and transcription, elements of bookkeeping and simple content and graphic production.

But the more immediate impact may not be the disappearance of entire professions. It could instead be the automation of individual tasks within those professions.

A call-centre worker, for instance, may increasingly rely on AI to handle routine queries while dealing personally with complicated cases. An accountant may use AI to process information while retaining responsibility for judgement and compliance.

The same applies across the wider economy.

In agriculture, AI can help farmers identify pests, assess weather risks and determine planting periods. It cannot, however, plough a farm, negotiate with buyers or resolve a dispute over land.

In healthcare, AI can support diagnosis and triage, but doctors, nurses and community health workers remain essential for physical care, trust and human judgement.

“The real risk in Kenya isn’t mass unemployment — it’s a skills mismatch,” Turuthi says.

The question facing workers, therefore, may be less about whether AI will replace them and more about whether they can learn to work alongside it.

“The practical question isn’t, will AI take my job, but, am I upskilling to work alongside it?” he says.

Kenya’s AI ambition

Kenya has recognised that widespread consumption alone will not be enough.

The government launched the Kenya Artificial Intelligence Strategy 2025–2030 in March 2025, with three core pillars: AI digital infrastructure; data and AI governance; and AI research, innovation and commercialisation. The strategy also identifies talent development, investment, governance, ethics, equity and inclusion as key enablers.

The implementation roadmap sets a broader ambition of increasing AI’s contribution to GDP in priority sectors, developing a stronger AI workforce, improving datasets and attracting greater investment in AI.

The government has also said it is integrating AI and digital skills into education curricula and developing workforce upskilling programmes.

Turuthi says Kenya has some important foundations, including data centres capable of supporting AI workloads and a strong fibre-connectivity ecosystem.

“Honestly, right now, we only have data centres that are AI capable and a vibrant fibre connectivity ecosystem. The rest, not yet, but the ambition is real and the plan is credible,” he says.

But infrastructure is only one piece of the puzzle.

Kenya needs reliable electricity and connectivity, computing capacity, skilled engineers and researchers, quality datasets and institutions capable of deploying AI responsibly.

The country also faces challenges in developing AI systems that understand local languages and contexts.

Government officials have acknowledged a “linguistic blindspot” in global AI systems, while research on natural-language processing in Kenya has highlighted the underrepresentation of many indigenous languages in digital datasets and tools.

Without investment in local capability, Kenya risks becoming a permanent customer of technologies developed elsewhere.

Yet the country has advantages: a young population, widespread mobile-money use, a growing technology and start-up ecosystem and expanding digital infrastructure.

The question is whether those advantages can be converted into ownership and innovation.

“The real test won’t be the strategy itself, but whether the money, training programmes and infrastructure projects actually get delivered on schedule. That’s the story worth following over the next few years,” Turuthi says.

The risks beyond jobs

AI’s impact will extend beyond the workplace.

Technology is making it increasingly easy to generate convincing text, photographs, audio and video. That creates new opportunities for fraudsters and misinformation campaigns.

A person’s voice or image can be manipulated. A company executive can be impersonated. False information can be produced and distributed at a speed that was previously difficult to achieve.

AI did not create misinformation, but it has lowered the barriers to creating convincing synthetic content. This makes verification, media literacy and accountability increasingly important.

Privacy presents another challenge.

As Kenyans upload documents into AI systems, they may expose identification details, financial information, confidential business records or other sensitive material.

Kenya’s Data Protection Act provides the legal framework for processing personal data, with the Office of the Data Protection Commissioner responsible for regulation.

But businesses still need clear internal rules governing what employees can enter into AI systems and how that information is stored, processed and accessed.

The stakes become even higher when AI is used in recruitment, lending, insurance or access to essential services.

If an automated decision affects a person’s livelihood, ability to obtain credit or access to a service, there must be mechanisms for explaining and challenging that decision.

So, is Kenya ready?

Kenya has clearly moved beyond the question of whether AI matters.

Businesses are experimenting with it. Workers are incorporating it into their routines. Consumers are using it in extraordinary numbers. And the government has established a national framework intended to make Kenya a stronger participant in the global AI economy.

But adoption is not the same as readiness.

Kenya still needs deeper investment in skills, infrastructure, research and locally relevant data. Businesses must move beyond impressive demonstrations and prove that AI delivers measurable value.

And as the technology becomes embedded in everyday life, safeguards must keep pace.

“Kenya isn’t hyping AI. Kenyans are simply using it, in huge numbers, faster than the rest of the world. The businesses that will actually benefit are the ones treating AI as a change to how they operate, not just another app on the phone,” Turuthi says.

That may ultimately be the dividing line between revolution and hype.

AI is unlikely to transform Kenya in one dramatic moment.

But whether it becomes a genuine engine of economic transformation or another overhyped technological promise will depend not on how quickly Kenyans download and use AI tools, but on whether the country can build the skills, infrastructure, data and institutions needed to capture the value created by them.