ICTworks https://www.ictworks.org/ Expertise in sustainable ICT4D for the developing world Sun, 26 Jul 2026 07:35:44 +0000 en-US hourly 1 https://wordpress.org/?v=7.0.2 https://i0.wp.com/www.ictworks.org/wp-content/uploads/2020/03/inveneo-logo.png?fit=32%2C32&ssl=1 ICTworks https://www.ictworks.org/ 32 32 105658519 Updated: Every African Country’s National Digital Health Strategy in 2026 https://www.ictworks.org/updated-every-african-countrys-national-digital-health-strategy-in-2026/ https://www.ictworks.org/updated-every-african-countrys-national-digital-health-strategy-in-2026/#respond Thu, 23 Jul 2026 04:37:38 +0000 https://www.ictworks.org/?p=20327 When I first compiled the list of every African country’s national eHealth strategy or digital health policy back in 2020, the dominant instrument across the continent was a five-year national eHealth strategy. They were written to a WHO Global Observatory template, often expired, frequently undiscoverable, and almost never legally binding. Six years on, that picture […]

The post Updated: Every African Country’s National Digital Health Strategy in 2026 appeared first on ICTworks.

]]> digital health strategy

When I first compiled the list of every African country’s national eHealth strategy or digital health policy back in 2020, the dominant instrument across the continent was a five-year national eHealth strategy.

They were written to a WHO Global Observatory template, often expired, frequently undiscoverable, and almost never legally binding. Six years on, that picture has changed enough that the original list is misleading more than it is helpful.

Sign Up Now for more African digital health insights

The headline shift since 2019 is the kind of strategy in each country. The WHO-templated eHealth document of the 2010s is being replaced by something bigger, more integrated with national digital transformation agendas, and increasingly statutory rather than purely strategic.

  • Kenya’s Digital Health Act of 2023 is the clearest example: a binding statute that establishes a Digital Health Agency and a Comprehensive Integrated Digital Health Information System, functionally superseding a 2016 policy that on paper still has another five years to run.
  • Egypt launched its National Digital Health Strategy 2025-2029 in November 2025, anchored on the Universal Health Insurance digital architecture rather than an isolated MoH workstream.
  • Mozambique’s first dedicated Estratégia de Saúde Digital de Moçambique 2025-2034 runs ten years, signalling that ministries are now sequencing investment beyond the donor project horizon.

Below is the updated table of operative national digital health instruments for all 54 African Union member states, current as of July 2026. Where I can link to a primary source on a ministry, regulator, or WHO portal, I do. Where I cannot, the entry is marked [SOURCE NEEDED] and I am asking you to send me the document.

National Digital Health Policies

11 Countries Missing Primary Text

Algeria, Benin, Burundi, Guinea-Bissau (the digital transformation strategy specifically), Liberia, Libya, Mali, Mauritania, Niger, Senegal, and Togo all share the same failure mode.

A strategy is documented in secondary reporting (national press, donor coverage, WHO/UNDP project pages) but no primary government-hosted PDF was locatable in July 2026. Benin is the same plus the additional complication that no successor exists. Sudan and Eritrea carry primary sources for the integrated HSSP but no standalone digital strategy.

If you have a working link to any of these primary documents, please email me so I can update the list and credit you in the next revision.

What Changed Since 2020

The 2020 list reflected the WHO Global Observatory survey universe, where the question was whether a country had a “national eHealth strategy” filed against a templated definition.

Most of the entries were five-year strategies, often expired by the time I compiled them, with primary text on Google Drive because the ministries had not posted them. The 2026 picture is different in three concrete ways.

1. Dominant instrument has migrated.

The five-year eHealth strategy is being replaced by either a longer-horizon digital health strategy of seven to ten years, an integrated PNDS or NHSSP that absorbs digital priorities, or, in Kenya’s case, primary legislation.

At least sixteen of the operative instruments in this table are dated 2024 or later. The instruments adopted across Burkina Faso, Cameroon, Côte d’Ivoire, Mali, Mauritania, Niger, and Senegal between 2023 and 2025 share architectural DNA with each other (interoperability layer, EHR, telemedicine, governance pillar, costed implementation) more than any of them visibly share with the Africa CDC Digital Transformation Strategy 2023-2030.

That is the clearest measure of how fast this moved.

When Bry Sylla and colleagues ran the same census against a stricter definition, counting only standalone digital health strategies developed after the 2020 WHO global strategy and still valid, they found eleven for the entire continent as of 31 December 2023, and concluded that 43 of 54 countries had no current post-2020 plan.

My sixteen and their eleven are not the same quantity measured twice. They excluded the integrated PNDS and NHSSP documents carrying digital chapters, which I count, and their cutoff predates most of what changed. Read together, the direction is not in doubt.

The continental coordinating frame exists. The policy evidence does not yet support claims that it is materially driving national alignment. The convergence is also shallower than it looks from the outside.

Sylla and colleagues found the eleven plans agreeing on vocabulary, with universal health coverage appearing in eight visions and quality of care in seven, while splitting three ways on sequencing logic:

  • build the institutional environment first (Cameroon, Namibia, Tunisia, Botswana),
  • deploy interventions and data infrastructure first (Guinea, Burundi, DRC, Malawi),
  • attempt both at once (Ghana, Ethiopia, Zambia).

Shared vocabulary, incompatible theories of change. Continental alignment does not produce that pattern.

2. The lusophone laggards have closed the gap.

In 2020 I had Angola at a 2010 strategy and Mozambique at a 2009-2014 one. Today Mozambique has its first dedicated 10-year digital health strategy, Angola is finalizing an HIS Strategic Plan with WHO and EU support, and Cabo Verde and São Tomé are mid-process.

3. Post-coup governments adopted new policies.

Burkina Faso’s PSSN 2025-2029 was signed in January 2025. Mali’s PSNSN 2024-2028 was developed with WHO, UNDP, World Bank, USAID, and UNICEF support. Niger validated its Schéma Directeur in May 2025 despite the post-coup environment. Health digitization apparently sits below the political fault line. That is a finding worth absorbing before assuming political instability automatically means digital health stagnation.

Key Weakness: Thin Documents, Not Just Missing Ones

Several of the post-2023 strategies were drafted with donor consultancy money under WHO or World Bank, validated at a workshop, and then never surfaced as a downloadable PDF on the ministry website that nominally owns them. The Algeria, Mali, Mauritania, Niger, and Senegal entries all carry [SOURCE NEEDED] tags for exactly that reason.

A national digital health strategy that exists only in the consultant’s deliverable and the launch press release is a different artifact than one whose legal text and budget annexes the ministry treats as governing.

Until recently that was where my argument stopped, because nobody had systematically read the plans we can get hold of. In December 2025, Sylla and colleagues did, assessing eleven plans against a grid combining the Walt-Gilson policy triangle with WHO and ITU planning standards. Their finding starts earlier than implementation and is harder to argue with.

Only Guinea’s plan reported a complete set of socio-economic and health context indicators. Ethiopia, Ghana, and Burundi reported none at all. More than half of the eleven never stated the health system challenges that the digital investment was supposed to solve. The enabling-environment pillars most often missing were workforce, legal frameworks, financing, and interoperability, which are the four things that determine whether any of this gets built.

Across the eleven plans, 148 digital health interventions are specified: 45% percent target healthcare providers, 25% target data services, 24% target system managers, 7% target clients.

  • Namibia’s plan contains two interventions in total. Botswana’s contains four. These are strategies written for ministries and clinicians, in countries whose stated vision in eight of eleven cases is universal health coverage, with patients as a rounding error.
  • Cameroon evaluated its 2020-2024 plan in 2025 before replacing it and scored system maturity at 1.8 out of 5. That is what honest evaluation looks like. Cameroon is not an outlier for being weak, only for admitting it. Cameroon ranked every component of its enabling environment at level 2, “developing.”
  • Ethiopia, the only other country publishing maturity scores, ranged from 2.29 on infrastructure to 3.37 on workforce. Two countries scored themselves against a standard tool. Both landed in the developing band. Nine did not score themselves at all.

One finding deserves the attention of anyone currently selling artificial intelligence into an African health ministry. None of the eleven plans mention AI, the Internet of Medical Things, blockchain, or drones. Not one.

The authors note this may partly reflect countries handling AI in separate national strategies, and Tunisia’s newly announced centre of excellence for AI in health suggests exactly that. The narrower claim : the instruments that actually govern health system digitization say nothing about the technology dominating every donor conversation about them.

Sub-Saharan Africa hosts 738 documented digital health interventions with massive duplication and limited coordination, and a strategy on paper has rarely been the binding constraint on that fragmentation. The continent has moved from “we cannot find the strategy” toward “the strategy exists but the ministry cannot show us the document.” That is real progress. It is not yet implementation, and among the eleven plans we can actually read, it is not yet good planning.

Please Help Improve This List

Please email me the primary government source for any of the 11 [SOURCE NEEDED] strategies. Send me the URL on the ministry website, the Africa CDC focal point name, or the consultancy that delivered the document. If you work in any of the ministries that adopted a strategy between 2023 and 2025 and the document is not online, ask why.

Two shortcuts I should have used earlier.

Neither is complete, which is the entire problem, but both beat searching ministry sites blind.

The 2020 list came together because readers pointed out what I had missed. The 2026 list will be useful only if you do the same. Which African country’s digital health strategy can you help us find?

The post Updated: Every African Country’s National Digital Health Strategy in 2026 appeared first on ICTworks.

]]> https://www.ictworks.org/updated-every-african-countrys-national-digital-health-strategy-in-2026/feed/ 0 20327 Meta Wants to Charge for WhatsApp Messages: Chatbot Builders Beware https://www.ictworks.org/meta-wants-to-charge-for-whatsapp-messages/ https://www.ictworks.org/meta-wants-to-charge-for-whatsapp-messages/#comments Tue, 21 Jul 2026 04:42:34 +0000 https://www.ictworks.org/?p=20468 Starting October 1 2026, anyone using WhatsApp as a platform to reach their target audience will be charged by Meta for outbound messages. Until that date, it will have been free to do so within a 24-hour window opened by the user’s last message, meaning that as long as your chatbot (or human agent) was […]

The post Meta Wants to Charge for WhatsApp Messages: Chatbot Builders Beware appeared first on ICTworks.

]]> whatsapp charges

Starting October 1 2026, anyone using WhatsApp as a platform to reach their target audience will be charged by Meta for outbound messages.

Until that date, it will have been free to do so within a 24-hour window opened by the user’s last message, meaning that as long as your chatbot (or human agent) was engaging and helpful enough, users seeking support for health, livelihoods, education and psycho-social support could chat for as long as they wanted to (and could afford to themselves).

All implementers had to pay for was the cost of the WhatsApp Business Service Provider (BSP) platform, such as Turn.io, as well as any re-engagement messages to wake up ‘dormant’ users. In the near future, any message sent in response to users will be charged at the same rate as ‘utility templates’, whose costs vary region by region.

This change has huge scale, sustainability, user experience and impact implications for many social impact organisations using, or thinking about using, WhatsApp to reach people across the Global South.

WhatsApp was the channel of choice

WhatsApp became the digital intervention channel of choice starting in 2018, when its API was released, allowing many of us to seize the opportunity to reach our target audience on the channels they actually use.

I was one of very few people designing chatbots back then, experimenting with chatbot copy-writing, design conventions, and figuring out the best way to leverage a fairly restrictive format into a rich experience.

Sure, the ‘chatbots’ we created were pretty basic, relying mostly on text, pre-determined decision-trees, and of course, emojis. But they allowed even small NGOs to reach high volumes of users with potentially life-changing information, advice, and support, on an app that was often zero-rated by MNOs, much less prone to connectivity issues than mobile web, and so embedded in users’ lives that it was less likely to be deleted from their phones to preserve memory or battery power.

Crucially, users trusted WhatsApp chatbots, loving the human-not-human chat-based format.

The pandemic turbo-charged the use of WhatsApp by development and humanitarian practitioners, including multilaterals like WHO who deployed chatbots on WhatsApp and other messaging apps to millions of users in over 19 languages, with the help of Turn.io and organisations like Reach Digital Health.

At a smaller scale, civil society organisations like the International Fact Checking Network, created a chatbot allowing users to submit health misinformation and get it fact checked in minutes. Today, WhatsApp usage represents 69% of all internet usage worldwide, and 83% of monthly users open it every day.

WhatsApp chatbots now feel like old news in the development space ( although we’re still trying to figure out how to measure their impact.)

GenAI chatbots talk more; Meta wants them to talk less

Whilst launching a chatbot for social impact no longer feels like a novelty, WhatsApp chatbot fever is in no-way diminished. GenAI means that we can finally create *actually* smart chatbots, either run entirely using a Large Language Model, or which include genA-powered components to handle specific conversations requiring more personalisation and flexibility.

What this means is that at the exact moment that we are being put under pressure to use genAI to make chatbot experiences richer, more personalised, more flexible, and more accurate, Meta’s new pricing model is forcing us to do the opposite.

Turn’s advice to customers using their platform is to ‘reduce the amount of back-and-forth and get users to what they’re looking for as quickly as possible.’

They position this as a user experience improvement as well as a cost-reduction method, and in a general sense, they’re not wrong – many chatbots could do better at being succinct and reducing their feature bloat.

But it’s also at odds with the parallel push towards AI-powered, conversational experiences – whose underlying agents will now have to be prompted to be ‘less chatty’, when a big reason for their integration was exactly this.

What this will cost implementers

Those with existing WhatsApp services are scrambling to calculate the implications. Informal estimates circulating among implementers range from $500 extra per month, to $4,500 per month to engage with around 10,000 users a month.

As one implementer shared: “this is a major constraint to scaling plans and sustainability projections.” For smaller NGOs, what was a cost-effective way of reaching high-volumes of individuals in their target audience with valuable messaging and support, will become prohibitively expensive.

Monthly Increase User Scale Country Use Case
$4,560 10,000 monthly users South Africa Sexuality & mental health
$3,616 3 million monthly interactions South Africa Learning & Teaching
$2,245 18,000 monthly users Argentina Mental health

On the other hand, because exact costs vary per country, some implementers are less concerned. Similarly, for chatbots where the service offering is less dense, and less conversational, the impact may be less significant.

As one implementer shared: “I am personally unconcerned. This is because in India, our charges are going from 2 US cents to 7 US cents. Our per-participant cost overall will remain nothing for us.” (their emphases).

Additionally, where users come from in the first place matters. Anyone who relies on Meta ads which click-to-WhatsApp for the majority of their marketing will still be able to message users for free within a 72 hour window of the first conversation started.

Given that average user lifetime is often only a day, those relying on Meta ads may not be affected at all: basically, don’t panic until you’ve done the maths for your particular set-up.

The human cost

Perhaps more seriously, this pricing change may also impact on the role of humans in increasingly automated services. WhatsApp is also used to support human-staffed Helpdesks, with trained operators able to jump in and out of an automated conversation to provide human-led support, including last-mile service provision referrals.

These agents are central to safeguarding processes, with users expressing suicidal ideation, survivors of Gender Based Violence, or women going through miscarriages referred as soon as possible to human support. Let’s imagine for a second how it would feel to tentatively disclose spousal abuse, or the loss of a baby, when our interlocutor wants to wrap up the conversation as quickly as possible for cost reasons.

How we can push back

As we speak, implementers leveraging WhatsApp, and the integration platforms who allow them to do so, are discussing ways to fight back, for example by:

  • Demanding clarity and seeking exemptions for non-profits from Meta;
  • Forming a coalition;
  • Engaging with country-level regulators;
  • Convincing donors to supplement budgets to cover the unexpected surplus.

Whether or not they prevail remains to be seen – the social sector has sought for years with limited success to get Meta to make changes that would enable us to use WhatsApp more effectively in the contexts in which we work.

Our collective bargaining power also ebbs away with every cut to international aid. In the meantime, here are some things you can do, right now.

This isn’t the end for WhatsApp chatbots, but it’s yet another lesson about relying on corporate behemoths to run essential services. Something to think about before you make GPT, Llama, Gemini or Claude central to your entire intervention – but you know that already.

Given their substantial profitability challenges to date, I would bet that the same thing will happen with genAI before too long.

By Isabelle Amazon-Brown who designs chatbots & AI for positive social impact

The post Meta Wants to Charge for WhatsApp Messages: Chatbot Builders Beware appeared first on ICTworks.

]]> https://www.ictworks.org/meta-wants-to-charge-for-whatsapp-messages/feed/ 1 20468 Apply Now: Google DeepMind AI for the Planet Accelerator https://www.ictworks.org/apply-now-google-deepmind-ai-for-the-planet-accelerator/ https://www.ictworks.org/apply-now-google-deepmind-ai-for-the-planet-accelerator/#respond Mon, 20 Jul 2026 04:12:33 +0000 https://www.ictworks.org/?p=20479 Asia-Pacific is the most disaster-prone region in the world. The UN Economic and Social Commission for Asia and the Pacific projects that climate-induced disaster losses could near US$1 trillion annually, or 3 percent of regional GDP, under a 2°C warming scenario. Green technology is gaining traction across the region, but it is not scaling fast […]

The post Apply Now: Google DeepMind AI for the Planet Accelerator appeared first on ICTworks.

]]> google deep mind funding

Asia-Pacific is the most disaster-prone region in the world. The UN Economic and Social Commission for Asia and the Pacific projects that climate-induced disaster losses could near US$1 trillion annually, or 3 percent of regional GDP, under a 2°C warming scenario. Green technology is gaining traction across the region, but it is not scaling fast enough to match the threat.

Google DeepMind AI for the Planet

The inaugural AI for the Planet accelerator from Google DeepMind, the AI research arm of Google, wants to close that gap. It will select 10 to 15 startups, research teams, and non-profits headquartered in Asia-Pacific that are applying frontier AI to environmental challenges.

This is three-month program offers selected organizations:

  • Access to the Google AI stack, including specialized frontier models
  • Dedicated mentoring from Google teams and AI experts
  • A week-long in-person bootcamp and three months virtual support
  • Potential cloud credits and free Cloud TPUs
  • A Demo Day to showcase solutions to investors and industry partners
  • Google Accelerator Alumni network of more than 2,000 startups and non-profits

Apply if your project focuses on nature, climate, agriculture, sustainability, or energy, with particular interest in AI-enabled nature protection, sustainable agriculture, and forest protection.

Applicants need a functional prototype or minimum viable product with early validation, an in-house technical team with deep AI and machine learning expertise, and a solution where AI is the core driver.

You must also show clear potential to integrate Google AI models such as Gemini, Gemma, AlphaEarth, Forestry, or Perch. At least two or three key leaders, including founders or CxOs, must actively participate.

Apply Now! Deadline is July 26, 2026

More Funding Opportunities

Please sign up now to get our email updates. Learn how to get startup funding for your technology business, and find new funding opportunities with donors.

We are constantly publishing fundraising guidance and competitive insights like this:

The post Apply Now: Google DeepMind AI for the Planet Accelerator appeared first on ICTworks.

]]> https://www.ictworks.org/apply-now-google-deepmind-ai-for-the-planet-accelerator/feed/ 0 20479 How I Built an Open Source Digital Marketplace for Women in Costa Rica https://www.ictworks.org/open-source-digital-marketplace-for-women-in-costa-rica/ https://www.ictworks.org/open-source-digital-marketplace-for-women-in-costa-rica/#comments Thu, 16 Jul 2026 04:43:11 +0000 https://www.ictworks.org/?p=20360 Honest ICT4D case studies are the ones that admit which technology choices were forced on us by the people we were trying to serve. I was the technical lead at UN Women HQ on Hecho por Mujeres, Costa Rica’s national e-commerce marketplace for women entrepreneurs living in poverty. What follows is the practitioner-level account of […]

The post How I Built an Open Source Digital Marketplace for Women in Costa Rica appeared first on ICTworks.

]]> e-commerce UN women

Honest ICT4D case studies are the ones that admit which technology choices were forced on us by the people we were trying to serve.

I was the technical lead at UN Women HQ on Hecho por Mujeres, Costa Rica’s national e-commerce marketplace for women entrepreneurs living in poverty. What follows is the practitioner-level account of what the constraints demanded, including the parts that most project reports leave out.

We Adapted, We Did Not Invent

Hecho por Mujeres is the Costa Rica deployment of UN Women’s Buy from Women platform, an open-source, cloud-based e-commerce system.

The platform launched on November 22, 2021 with 54 women entrepreneurs, five public institutions, and four UN agencies. Today it serves 134 registered women sellers. The joint programme is implemented by UN Women, UNDP, ILO, and FAO, with IMAS as lead national counterpart and FIDEIMAS as the long-term administrator.

The sellers are women already enrolled in IMAS, FIDEIMAS, and INAMU social programs. The marketplace did not recruit entrepreneurs. It built a digital storefront for women the Costa Rican state had already identified as living in poverty.

Magento Was a Budget Decision

The platform runs on Magento Open Source. The reason was simple: money. Every dollar saved on proprietary licensing went directly to seller onboarding, training, and capacity building. For a UN-funded programme, that trade-off was obvious.

But open source comes with a catch. Open source means public code, and public code means a public attack surface. Anyone can read the codebase, which means anyone can look for vulnerabilities. Magento is one of the most targeted e-commerce platforms in the world for exactly this reason.

We took that seriously from day one. To keep the platform secure we implemented the latest Magento security patches at launch and maintained a strict patching schedule tied to Adobe’s security bulletins. We added two-factor authentication on all staff accounts, role-based access controls to separate seller, buyer, and admin environments, and hardened admin URL paths to reduce exposure.

The women selling on this platform are economically vulnerable. Their home addresses and financial details live in this system. A security breach is a real harm to real people. That shaped every decision we made.

Inclusion Bottleneck: Payment Infrastructure

The seller cohort does not have bank accounts. That single fact made every off-the shelf payment assumption wrong.

We integrated GreenPay because it offered three things the women needed and one thing the platform needed. The women needed card acceptance without a merchant account, payment links they could send over WhatsApp, and Spanish-language support they could access.

The platform needed PCI-DSS-compliant tokenization so card details were captured and stored by GreenPay, not by Hecho por Mujeres. The platform itself never holds sensitive payment information. That is the architecture to responsibly handle the data of women whose financial security has no cushion.

One clarification: GreenPay is not Costa Rican mobile money in the way M-Pesa is Kenyan mobile money. Costa Rica’s domestic real-time payment system is SINPE Móvil. GreenPay is a card and digital payment gateway.

The inclusion problem it solved was that a seller without a bank account could receive payment from a buyer with a card, and the platform is the regulated intermediary. That is bridging women into existing commerce infrastructure, not leapfrogging banking entirely.

Last-Mile Logistics Came From the Post Office

Costa Rica has something most ICT4D markets do not: a functioning national postal service. Correos de Costa Rica operates 110 branches nationwide and runs a dedicated e-commerce logistics platform through its EMS service. We integrated Correos rather than building custom logistics.

A seller in a remote canton now ships at the same rates and on the same timeline as a commercial business in San José.

The lesson is that the absence of usable national logistics infrastructure is the silent veto on rural e-commerce in most of the world. This fact rarely surfaces in project planning until the platform launches and orders cannot be fulfilled.

Sustainability Was an Institutional Question

The most important decision on this project had nothing to do with Magento or GreenPay or Correos. It was that FIDEIMAS was the platform’s owner from before launch, not after.

The default ICT4D pattern is the opposite.

  • A donor funds a pilot.
  • An implementing partner builds a system.
  • The system runs on a donor budget for three years.
  • At the end of the cycle, there is no handover plan.
  • The platform either dies or is bolted onto a ministry that did not ask for it.

In Costa Rica we did it the other way.

FIDEIMAS was the implementing counterpart from project design. The platform’s seller intake form lives on the FIDEIMAS website. The classification commission is led by FIDEIMAS staff. In March 2022, the Ministry of Economy, IMAS, INAMU, and INA signed a framework cooperation agreement to formalise the interagency commitment.

The marketplace does not collapse when UN funding ends because UN funding was never the operational base.

What to Learn From My Experience

Five things, and none of them about technology preferences:

  1. Publish your security operations record. UN Women and implementing partners should publish the full patching cadence, incident logs, and data minimization decisions for every Buy from Women deployment. The platform is open source. Its security posture should be too.
  2. Budget security at 10–15% of platform cost from year one. Anything less offloads risk onto the most vulnerable users.
  3. Stop calling open-source platforms cheap. They are cheaper at procurement and more expensive at sustainment. That is a fair trade only if the sustainment budget is real.
  4. Demand institutional ownership from project design. The Costa Rica sequence is the replicable model.
  5. Stop publishing case studies that hide the constraints. The architecture of Hecho por Mujeres reads coherently because we accepted that the inclusion problem dictated the technology. Most platforms that fail in this space did the reverse.

The technology choices were the easy part. The honest part was admitting which choices were not ours to make.

By Itban Omair, Manager ERP Applications, UN Women HQ New York 

The post How I Built an Open Source Digital Marketplace for Women in Costa Rica appeared first on ICTworks.

]]> https://www.ictworks.org/open-source-digital-marketplace-for-women-in-costa-rica/feed/ 1 20360 Wow! Open Source Costs NGOs More Than Commercial Software https://www.ictworks.org/open-source-software-costs-ngos/ https://www.ictworks.org/open-source-software-costs-ngos/#comments Tue, 14 Jul 2026 04:07:31 +0000 https://www.ictworks.org/?p=20301 For 15 years, open source has functioned as something close to scripture in ICT4D. Principles for Digital Development enshrined it as Principle 6. Digital Public Goods Alliance built a global movement around it. UNICEF Innovation declared that it would only build open source in 2026 Donor procurement guidance treats “we’ll open source it” as proof […]

The post Wow! Open Source Costs NGOs More Than Commercial Software appeared first on ICTworks.

]]> open source software

For 15 years, open source has functioned as something close to scripture in ICT4D.

  • Principles for Digital Development enshrined it as Principle 6.
  • Digital Public Goods Alliance built a global movement around it.
  • UNICEF Innovation declared that it would only build open source in 2026
  • Donor procurement guidance treats “we’ll open source it” as proof of sustainability.

Anyone who has tried to defend a proprietary tool in a proposal review knows the social cost of confronting Open Source software true believers.

So I paid attention when NetHope published Harnessing AI for Humanitarian Impact – a synthesis of 11 humanitarian AI deployments at IRC, Mercy Corps, ICRC, GiveDirectly, and others. Buried in the cross-cutting analysis is a sentence that quietly takes a hammer to one of our sector’s foundational beliefs.

Here it is, verbatim from the report:

  • “Open-source does not eliminate cost, it redistributes it. Licensing fees are replaced by coordination demands: maintaining codebases, documenting tools, onboarding contributors, and governing shared infrastructure. For resource-constrained NGOs, these coordination costs can equal or exceed the price of a commercial license.”

Read that again. Seriously. Its a bold statement that deserves a moment’s reflection.

NetHope, an organization that has spent two decades helping NGOs adopt technology, is telling us that the total cost of ownership math we have been using to justify open source mandates may be wrong for the typical operational humanitarian organization.

Sign Up Now to understand Open Source TOC

Before You Quote This Report

I am going to give you the contrarian argument that this finding deserves. But first, the disclosure that the report itself does not foreground.

The NetHope Center for the Digital Nonprofit, which produced this analysis, lists four partners on the inside cover: Box, Okta, PagerDuty, and Microsoft.

  • Box sells proprietary cloud storage.
  • Okta sells proprietary identity management.
  • PagerDuty sells proprietary incident response.
  • Microsoft does not need an introduction.

Every funder of this report has commercial reasons to want NGOs spending less time on open source and more time on commercial licenses.

This does not make the report’s findings wrong.

NetHope is a serious organization, the case studies are well-documented, and the open source critique is one paragraph in a 13-page analysis that also praises fAIr and POMELO for publishing their code on GitHub.

But when a research product funded by proprietary vendors has a finding that proprietary software may be cheaper than open source for NGOs, intellectual honesty requires naming the conflict before citing the conclusion.

Hold both things in mind. The funding source is a real reason for skepticism. The argument may still be correct.

The Argument Is Probably Correct Anyway

When NetHope’s vendor-funded report says coordination costs can equal or exceed commercial licensing, we should listen because they are echoing what four different organizations have already said with their own money on the line.

1. OpenLMIS

The OpenLMIS community came to almost the same conclusion six years ago, and they were funded by USAID and the Gates Foundation, not Microsoft. When OpenLMIS went looking for sustainability, they ended up with three options:

  1. Charge implementations for ongoing maintenance,
  2. Expand to private clinics as paying customers,
  3. Hand core software maintenance to a private partner.

They chose option three and partnered with a commercial firm called Vitalliance in 2021. The “free” open source supply chain tool was kept alive by a private company that needed it integrated into its commercial value chain.

2. DEEP

DEEP, the humanitarian data analysis platform documented in NetHope’s report, did not get that lucky. It shut down in 2024 not because it failed but because its funding cycle ended. Open source did not save it. Nobody was paying the coordination tax.

3. Norwegian Red Cross

Mikael Hailu of the Norwegian Red Cross made the same point in August 2025, warning that most Global South open source tools are “mission-specific, tightly bound to domains like public health, education, or social services” with narrower user bases than the Apache or Linux comparisons we love to invoke.

Foundation-backed open source works for Kubernetes because Google, IBM, and Red Hat assign full-time engineers to it. Nobody is assigning full-time engineers to your mHealth tool in Malawi.

4. Development Gateway

Development Gateway argued bluntly last year that “open source is seen as an inherently sustainable option. This is a misunderstanding of both the value of open source and technology lifecycles.”

They had watched an internationally-applauded system go offline because the government did not budget for server costs. Open source did not prevent that. It may have made it more likely, by allowing everyone in the funding chain to assume someone else was carrying the maintenance load.

What This Means for Software Procurement

Three things should change in how the sector talks about open source in 2026, and none of them require abandoning the principle.

First, stop letting “we will open source it” function as a sustainability plan.

It is not one. A real sustainability plan names who pays for coordination, maintenance, retraining, hosting, security patching, and contributor governance for the next five years, in dollars, with a named funder. If your software development proposal does not have that, the open source clause is decoration.

Second, budget the open source coordination layer explicitly as a line item.

NetHope’s framing of open source as a “cost-reallocation strategy” rather than a cost-reduction strategy is the right mental model. The cost did not vanish. It moved from a vendor invoice to your staff calendar, where it is harder to see and easier to underfund.

Third, stop treating commercial vendors as a moral failure. 

Sometimes a managed commercial software services with a service-level agreement, a security team, and a contractual obligation to maintain the product is the cheaper, more sustainable, more accountable choice for a particular humanitarian use case.

The open source maximalist position, that NGOs funded with public money should never use proprietary tools, has produced real casualties: tools that died because nobody was paid to keep them alive, and procurement decisions made on ideology rather than fit.

Open Source Is Great – In Specific Cases

Open source remains essential for digital public infrastructure, for tools that need to outlive any single vendor, for code that should not be held hostage to one company’s pricing decisions.

fAIr and POMELO publishing on GitHub is genuinely good. The community-generated training data fAIr uses to correct Global North bias is exactly the kind of public good the sector should be building.

But for the operational software tools sitting in front of an NGO right now, the question is no longer “open or proprietary.” It is “who is funding the coordination layer, and for how long.” The honest answer to that question will determine whether the tool exists in 2030. The license attached to it will not.

The post Wow! Open Source Costs NGOs More Than Commercial Software appeared first on ICTworks.

]]> https://www.ictworks.org/open-source-software-costs-ngos/feed/ 2 20301 Please Stop Designing Humanitarian Solutions for Donor Funding https://www.ictworks.org/designing-humanitarian-solutions-for-donor-funding/ https://www.ictworks.org/designing-humanitarian-solutions-for-donor-funding/#respond Thu, 09 Jul 2026 04:08:21 +0000 https://www.ictworks.org/?p=20302 If you visit DEEP on the Wayback Machine, you will not find the AI-powered humanitarian analysis platform that UN OCHA, UNHCR, UNICEF, and the Internal Displacement Monitoring Centre had built for their secondary data review workflows around for nearly a decade. You will find this: “The DEEP was a trusted humanitarian analysis platform that fostered […]

The post Please Stop Designing Humanitarian Solutions for Donor Funding appeared first on ICTworks.

]]> deep digital solution

If you visit DEEP on the Wayback Machine, you will not find the AI-powered humanitarian analysis platform that UN OCHA, UNHCR, UNICEF, and the Internal Displacement Monitoring Centre had built for their secondary data review workflows around for nearly a decade.

You will find this: “The DEEP was a trusted humanitarian analysis platform that fostered collaborative and innovative solutions. Due to funding cuts, it is no longer available, but its legacy will continue to live on.”

That is what failure looks like in our sector in 2026. A tool that worked. Funding that did not work.

DEEP, the Data Entry and Exploration Platform, is one of the 11 case studies documented in NetHope’s new Harnessing AI for Humanitarian Impact report. It is also the only one in the report that no longer exists. The report’s framing of why it died deserves to be quoted in full because it indicts something much larger than DEEP:

  • “discontinued not because it failed but because its funding did.”

Read that sentence again with the rest of the report’s case studies in mind.

  • CRS dashboards needing maintenance budgets that don’t exist.
  • NRC self-hosting AI infrastructure to control costs.
  • PDC scaling to enterprise infrastructure without enterprise funding.

Every one of these tools is potentially the next DEEP. The only difference is timing.

Sign Up Now to advert disaster in your next project

Misdiagnosing Pilotitis for Years

The ICT4D community has spent years blaming pilotitis on the projects themselves. Bad design. Wrong context. Insufficient local ownership. No business model. Every conference panel on this topic eventually circles back to the same set of project-level remedies:

  • better needs assessments,
  • more co-design,
  • stronger M&E,
  • clearer theories of change.

DEEP punctures that diagnosis. By any measure ICT4D advocates have published over the last decade, DEEP was the success story. It had everything we keep saying makes pilots succeed.

  • Multi-stakeholder governance with OCHA, UNHCR, UNICEF, and IDMC.
  • Real users running real workflows.
  • Documented operational impact across multiple humanitarian responses.
  • Open source code so no single vendor could hold it hostage.
  • Eight years of iterative development.

It still got switched off.

The Digital Impact Alliance calls the four-to-five-year mark the “funding valley of death.” DEEP made it past year eight and still fell in. NetHope’s framing is the sharpest line in the report:

  • “The sector is using project-cycle grants to build what functions as long-term digital public infrastructure, financing roads like meals, expecting one-time investment to sustain something that requires perpetual upkeep.”

The diagnosis is structural, not behavioral. You cannot fix it with a better logframe.

Donor Funding Is Fickle. Government Is Not.

Most NGO leaders know that donor funding is structured to be fickle.

Foundations rotate strategic priorities every five to seven years. Bilateral aid agencies change direction with each administration, as anyone who watched USAID’s death or FCDO’s cuts can confirm. Even committed donors operate on grant cycles that rarely exceed three years, because that is how their own boards measure them.

This is not a flaw in donor behavior. It is the design specification.

Donors are supposed to take catalytic risks, fund innovation, and exit. They are not supposed to operate digital infrastructure in perpetuity. When we build humanitarian AI tools that depend on continuous donor funding to survive, we are asking donors to do a job they were never structured to do.

Government funding works differently. National budgets are annual but recurrent.

Once a tool is integrated into a ministry’s line items, it persists across administrations because civil service continuity carries it.

Design for Government Funding

Most humanitarian AI tools cannot follow the DHIS2 path because they operate across borders, in conflict zones, or in spaces where no single government has authority.

  • A flood prediction model for the Sahel cannot be owned by one ministry.
  • A chatbot serving displaced Rohingya cannot depend on the Bangladeshi government’s annual budget.

The DEEP problem is harder than the DHIS2 problem. But “harder” is not the same as “impossible.” The shift starts with three concrete design choices, all of which need to happen at the proposal stage, not at the year-three sustainability review.

1. Identify the government procurement pathway before you write the technical specification.

Our recent analysis of why digital health pilots fail at scale made this point bluntly: the real challenge isn’t proving your intervention works. It is designing something the government can procure within existing systems.

For multi-country humanitarian tools, this means identifying which national agencies, regional bodies, or UN member-state contributions could plausibly absorb operational costs once the donor exits, and designing the tool to fit those procurement vehicles from day one.

2. Build the tool to integrate into existing government infrastructure rather than parallel to it.

fAIr’s mapping outputs feed OpenStreetMap, which governments already use. GiveDirectly’s flood triggers could feed national disaster management agency workflows. The tools that survive donor exit are the ones that have made themselves indispensable to systems governments are already paying to operate.

3. Treat donor funding as the catalytic phase, not the operational phase.

NetHope’s recommendation that funders shift to multi-year operational endowments is necessary but not sufficient. Even multi-year endowments end.

The sustainable funding plan names a non-donor entity (a government, a regional body, a recurring user-fee structure, a private-sector integration) that will carry the operational cost in year six and beyond.

If your proposal cannot name that entity, you are building DEEP again. A sustainable humanitarian AI tool is one that survives its donor exit. Everything else is just timing.

The post Please Stop Designing Humanitarian Solutions for Donor Funding appeared first on ICTworks.

]]> https://www.ictworks.org/designing-humanitarian-solutions-for-donor-funding/feed/ 0 20302 Time to Get Serious About AI in International Development https://www.ictworks.org/time-to-get-serious-about-ai-in-international-development/ https://www.ictworks.org/time-to-get-serious-about-ai-in-international-development/#respond Tue, 07 Jul 2026 04:11:21 +0000 https://www.ictworks.org/?p=20448 For three years, “AI for development” has been a conversation about possibility. Capability demonstrations. Pilot dashboards. Workshop slide decks. A thousand experiments in a thousand offices, each producing one good output, each one ending with the same question: what would it take to do this at scale, and to do it in a way the […]

The post Time to Get Serious About AI in International Development appeared first on ICTworks.

]]> ai4dev barcelona

For three years, “AI for development” has been a conversation about possibility. Capability demonstrations. Pilot dashboards. Workshop slide decks.

A thousand experiments in a thousand offices, each producing one good output, each one ending with the same question: what would it take to do this at scale, and to do it in a way the sector can actually trust?

There wasn’t really a good answer to that question. Until now.

The conditions for serious AI in international development – workflows that scale, shared infrastructure, good governance with guardrails, accountability that communities can rely on – are within operational reach for the first time.

The technology is mature enough. The standards are emerging. The donor conversations have evolved. The institutions that need to coordinate are, increasingly, seeking each other out.

Please join this September’s ECOSYSTEM Summit in Barcelona to operate AI at scale.

Time to Get Serious

For AI in development, the technology is at a stage to be taken seriously. Structurally.

But for the sector to take proper advantage of the potential of AI and other technologies, this requires funders, frameworks and institutional leadership to be aligned.

Getting serious looks like:

  • An NGO that adopts AI workflows with the same change-management discipline it applies to a new financial-controls system.
  • A humanitarian organisation that uses AI to support frontline allocation decisions, not just to write the sitrep faster.
  • A donor that asks the same questions of every AI tool in its portfolio and funds the institutional capacity to answer them.
  • A research community that can deliver evidence at the speed operational decisions actually need.
  • A sector that asks whose reality is encoded into its data – because AI that misrepresents the people it serves is not accountable, however well-audited.

Today, each of these is within reach. None of them is universal. Closing the gap between possible and universal is the work ahead.

ECOSYSTEM Summit in Barcelona

Three discussions over three days at the ECOSYSTEM Summit, taking place in Barcelona, from 16-18 September.

Day 1: AI 4 Development

Day 1 of the three-day programme convenes development agencies, donors, multilaterals, NGOs, foundations, researchers, data specialists, and technology partners.

Voices include the Gates Foundation, UNDP/IATI, World Bank, ONE Campaign, Mercy Corps, OpenAI, Google, J-PAL, GIZ Data Lab, CARE, Grand Challenges Canada, WFP, FAO, Digital Green, 60 Decibels, BAO Systems, and AidData — with donor participation from NORAD, FCDO, GIZ, and Sida.

At the core of the AI4D programme is the question for the IATI community and the broader sector about how to realise data’s full potential: how can AI improve development decisions and delivery while reinforcing transparency, accountability, data quality, and local agency?

Day 2: Ownership Economy

Day 2 of the Summit takes up the questions of who owns the data, who owns the model, who owns the upside, and who carries the risk when something breaks. As AI reshapes data, incentives and institutional power, what governance models can keep opportunity broadly distributed?

Engage with specialists from UN DESA, the London School of Economics, Social Economy Europe, Fairbnb, Ecosia and more.

Day 3: Trusted Commerce

Day 3 asks what standards, shared data rails and verification systems are needed for sustainability, procurement and measurable impact? While a single organisation can operate responsibly on its own, a sector cannot do so at scale on the basis of dozens of trustworthy islands that do not interoperate.

Sessions from Earthshot, Bureau Veritas, Bluesign, CommonShare, Traceability Lab and others will demonstrate that building something together isn’t just possible…it’s already happening.

Where Tools and Tribes meets Trust

Join us at ECOSYSTEM – Norrsken House, Barcelona, 16-18 September 2026.

AI4D. Ownership Economy. Trusted Commerce. Creating cross-sector relationships that turn fragmented innovation into something that really works, at scale. And laying the foundation for how to build economic systems that are intelligent, equitable and accountable.

Participants can attend any of the days as a stand-alone day or stay for the full ECOSYSTEM programme. Early bird pricing closes on June 30th.

The post Time to Get Serious About AI in International Development appeared first on ICTworks.

]]> https://www.ictworks.org/time-to-get-serious-about-ai-in-international-development/feed/ 0 20448 Apply Now: $100,000 for African AI Startup Founders https://www.ictworks.org/apply-now-100000-for-african-ai-startup-founders/ https://www.ictworks.org/apply-now-100000-for-african-ai-startup-founders/#comments Mon, 06 Jul 2026 08:45:54 +0000 https://www.ictworks.org/?p=20461 African AI startups captured barely 1 to 1.5% of global AI spending in 2025, even as AI pulled in close to half of all venture dollars worldwide. At the earliest stages the gap is wider still: in the second quarter of 2025, startups across the continent raised just $14 million in AI deals, about 0.02% […]

The post Apply Now: $100,000 for African AI Startup Founders appeared first on ICTworks.

]]> mest ai funding africans

African AI startups captured barely 1 to 1.5% of global AI spending in 2025, even as AI pulled in close to half of all venture dollars worldwide. At the earliest stages the gap is wider still: in the second quarter of 2025, startups across the continent raised just $14 million in AI deals, about 0.02% of the $47.3 billion invested in AI globally.

The constraint is not talent. It is capital, and the pathways to turn technical skill into a fundable company.

$100,00 Pre-Seed Investments

The AI Startup Program from the Accra-based Meltwater Entrepreneurial School of Technology that has trained African software entrepreneurs since 2008, is open for the 2027 cohort, offering select teams up to $100,000 in pre-seed investment.

Apply if you are:

  • West or East African citizens, ages 21 to 35
  • Demonstrated software development experience and a basic understanding of AI
  • A bachelor’s degree, diploma, or equivalent hands-on experience

The cohort will get seven months of full-time, residential training in Accra, Ghana, followed by a four-month incubation phase for the strongest ventures. Participants form teams, validate an AI business idea, build a working product (an MVP, or minimum viable product), and learn from practitioners at OpenAI, Perplexity, Google, and Meltwater.

Apply Now! Deadline is July 20, 2026

More Funding Opportunities

Please sign up now to get our email updates. Learn how to get startup funding for your technology business, and find new funding opportunities with donors.

We are constantly publishing fundraising guidance and competitive insights like this:

The post Apply Now: $100,000 for African AI Startup Founders appeared first on ICTworks.

]]> https://www.ictworks.org/apply-now-100000-for-african-ai-startup-founders/feed/ 4 20461 Hidden Climate Crisis Destroying Software Developer Productivity https://www.ictworks.org/hidden-climate-crisis-destroying-software-developer-productivity/ https://www.ictworks.org/hidden-climate-crisis-destroying-software-developer-productivity/#comments Thu, 02 Jul 2026 04:48:46 +0000 https://www.ictworks.org/?p=20112 The development community is obsessing over the wrong climate threat. While everyone debates if data centers are causing rising temperatures, a groundbreaking study from Bangladesh reveals the real killer: heat doesn’t just slow down individuals. High temperatures destroy the collaborative advantage that makes teams twice as productive as solo workers. There are serious implications for […]

The post Hidden Climate Crisis Destroying Software Developer Productivity appeared first on ICTworks.

]]> bangladesh dhaka software developers

The development community is obsessing over the wrong climate threat.

While everyone debates if data centers are causing rising temperatures, a groundbreaking study from Bangladesh reveals the real killer: heat doesn’t just slow down individuals. High temperatures destroy the collaborative advantage that makes teams twice as productive as solo workers.

There are serious implications for ICT4D solutions.

As cognitive work increasingly shifts to South Asia, Southeast Asia, and Africa, we’re heading toward a collaboration crisis that could undermine the foundation of modern digital development work.

Sign Up Now for more climate crisis insights 

Team Productivity Collapses at 29°C

Researchers at UC San Diego conducted a rigorous field experiment with 232 computer programmers in Dhaka, Bangladesh, randomly assigning participants to work either individually or in pairs under two temperature conditions: a control room at 24°C and a warm room at 29°C.

The temperature differential was intentionally modest. It was well below the 30-33°C thresholds where previous studies found individual productivity impacts.

The results challenge everything we think we know about heat and work performance.

Individual software programmers showed no productivity decline in warm conditions. Teams in control conditions were nearly twice as likely as individuals to complete coding features successfully. But here’s the critical finding: teams in warm rooms performed no better than individuals working alone.

Heat completely eliminated the collaborative advantage.

Teams in control rooms were more than twice as likely to add features compared to teams in warm rooms. The 44 percentage point drop in team performance represents the destruction of teamwork’s core value proposition.

Why This Matters for Digital Development

The basic ICT4D concepts we teach emphasize collaboration and partnership as fundamental to successful technology implementation. Yet we’ve been blind to how rising temperatures systematically undermine these collaborative processes.

Consider the geography of digital development work.

Bangladesh’s IT sector, worth $1.4 billion and supporting 300,000 jobs, operates in conditions where average temperatures during the study’s “cool” season still reached 31.2°C. India’s tech hubs face similar conditions. Kenya’s Silicon Savannah regularly exceeds 29°C.

Meanwhile, digital skills development initiatives increasingly emphasize teamwork and collaborative problem-solving.

The timing couldn’t be worse. According to the World Bank’s latest climate and jobs analysis, the International Labour Organization projects that 2.4 billion workers (71% of the global workforce) are now exposed to excessive heat during work. Heat exposure caused 640 billion potential labor hours to be lost in 2024, with productivity losses equivalent to $1.09 trillion.

But these aggregate statistics miss the collaboration crisis.

The Bangladesh study reveals that the real damage isn’t distributed evenly. It’s concentrated precisely where modern economies create the most value: in teams.

Heat Destroys Diverse Teams First

The research uncovered an even more disturbing pattern.

Heat’s impact wasn’t uniform across all teams. It systematically targeted diversity. Mixed-gender teams in warm rooms were 55% less likely to complete features compared to same-gender teams in cool conditions. Teams with different levels of academic experience showed similar patterns.

Post-experiment surveys confirmed the mechanism: team members in warm conditions, particularly in diverse teams, rated their partners 37 percentage points lower and were 31 percentage points more likely to prefer different teammates for future work.

This finding has serious implications for ICT4D’s equity agenda.

Development organizations have invested heavily in promoting gender balance and diverse teams based on decades of research showing their superior problem-solving abilities. Heat stress systematically undermines these advantages, creating a climate penalty that falls disproportionately on the diversity we’ve worked to achieve.

Why High Temperatures Slow Productivity

Crucially, the study found no evidence that teams in warm rooms exerted less effort. They typed the same number of characters per minute and spent equivalent time actively working. The problem was coordination failure.

  • “Heat may be reducing team performance through mechanisms other than directly affecting individual performance, for example, by increasing coordination costs or the likelihood of coordination failure.”

High temperatures appear to slow communication, impair decision-making, and reduce the effectiveness of collaborative problem-solving.

Homogeneous teams adapted by taking more breaks to manage heat stress. But diverse teams couldn’t coordinate even these basic coping strategies effectively, taking fewer breaks despite worse performance outcomes.

The Broader Climate-Development Crisis

This research reveals just one facet of climate change’s assault on modern work structures. As the development sector increasingly relies on digital transformation approaches that emphasize collaboration, cross-functional teams, and partnership models, rising temperatures systematically erode these collaborative advantages.

The geographic concentration makes this crisis particularly acute.

The same regions experiencing rapid growth in cognitive industries also face the most severe temperature increases with the least climate control infrastructure.

We cannot continue treating climate adaptation as a separate issue from digital development effectiveness.

  • ICT4D program design should now include temperature impact assessments.
  • Organizational development strategies need heat stress protocols.
  • Partnerships must consider climate resilience for collaborative work.

The Bangladesh programmers working in 29°C conditions weren’t facing extreme heat by global standards. They were experiencing the new normal for billions of workers in regions where we’re scaling our most important digital development initiatives.

We need to adapt our collaboration models before climate change systematically destroys the teamwork advantages that modern development depends on.

The post Hidden Climate Crisis Destroying Software Developer Productivity appeared first on ICTworks.

]]> https://www.ictworks.org/hidden-climate-crisis-destroying-software-developer-productivity/feed/ 1 20112 Borrow the Ceiling While Building Africa’s Artificial Intelligence Floor https://www.ictworks.org/borrow-the-ceiling-while-building-africas-artificial-intelligence-floor/ https://www.ictworks.org/borrow-the-ceiling-while-building-africas-artificial-intelligence-floor/#respond Tue, 30 Jun 2026 04:46:21 +0000 https://www.ictworks.org/?p=20456 On June 9, an American company shipped its most capable artificial intelligence model. Three days later, a letter from the US Commerce Department forced it offline for every user on the planet, with no notice and no appeal. The model was Anthropic’s Fable 5. The mechanism was an export control directive that barred access by […]

The post Borrow the Ceiling While Building Africa’s Artificial Intelligence Floor appeared first on ICTworks.

]]> artificial intelligence for development in africa

On June 9, an American company shipped its most capable artificial intelligence model. Three days later, a letter from the US Commerce Department forced it offline for every user on the planet, with no notice and no appeal.

The model was Anthropic’s Fable 5. The mechanism was an export control directive that barred access by any foreign national, so the company disabled it worldwide to comply. Other models kept running. This one did not.

Whatever you make of the security rationale, which Anthropic itself disputes, the structural fact is plain. A government on another continent reached into a service that African ministries, startups, and clinics were beginning to build on, and switched it off. As Maathangi Mohan says, the plug was in someone else’s hand.

The dominant response to that problem treats it as an access gap. Close the “AI divide,” the argument goes, by building data centers and importing GPUs until Africans can reach the same frontier models everyone else uses. Access matters.

But access to a model a foreign regulator can revoke on a Friday afternoon is not capability. It is a subscription with a kill switch. The useful question is which layers of the stack African countries have to own so that someone else’s decision cannot take the floor out from under its public services.

Sign Up Now for more African AI ownership insights

Someone Else’s Model Is Not a Strategy

Start with the hardware, because everything sits on top of it. African countries hold less than 1% of the world’s data center capacity while housing roughly 18% of its people. The United States and China alone operate more than 90% of the world’s usable AI compute, and more than 150 countries, the whole continent among them, have almost none.

Most African data already sits on servers in Europe and North America, which means it sits under foreign law, including a US statute that lets American authorities compel disclosure from US companies wherever the servers physically are.

When capability lives in someone else’s data center, it can be withdrawn for someone else’s reasons: national security, trade leverage, a complaint filed by a rival, a change of administration. The people cut off are rarely the intended target.

An LMIC health worker whose screening tool runs through a frontier API is collateral in a contest between great powers and frontier labs. Dependence on a revocable service is a borrowed capability, and the lender can call the loan without notice and without recourse. Fable 5 was the demonstration. It will not be the last one.

“Made in Africa” is a List of Layers

This is where Varaidzo Matimba’s framing earns its keep. In her 2025 landscape study for the MERL Tech Initiative, she defines Made in Africa AI as systems that originate from, are accountable to, and serve African communities, rooted in methods that prioritize local definitions of success rather than imported metrics.

The question she keeps putting to practitioners: who holds the power to design, deploy, and interpret the system.

Read that way, ownership is not one decision. It is a stack of them: whose data trains the model, whose languages it speaks, which problems it is pointed at, who evaluates whether it works, where it runs, and who can switch it off.

Some of those layers Africa cannot realistically own this decade. Frontier pretraining is one, and chasing a national supercluster to get it is, as I have argued here before, mostly a way to sell Nvidia chips.

But most of the layers that decide whether a tool helps a Kenyan farmer or a Malawian nurse are exactly the layers that can be built and governed on the continent, and they are cheap by comparison.

Layers Worth Owning

Small, open, on-device models are not only cheaper, they are un-revocable. A model whose weights you have downloaded, fine-tuned on your own data, and deployed on a community health worker’s handset cannot be turned off by a directive in Washington.

There is no kill switch because there is no switch: no API in Virginia, no license server, no cloud account to suspend. The compute sits where the user already is, and so does the control. The escape from the Western AI trap is to fine-tune small, task-specific models and run them on the phones people already carry, at a fraction of the cost of the frontier.

Language is the clearest case. Africa is home to more than 2,000 languages, and a 2025 review of large and small language models found that only about 42 of them have any meaningful support, with more than 98% severely underrepresented or ignored and just three scripts broadly covered.

No amount of imported compute fixes a gap that exists because the data, the tokenizers, and the evaluation were never built for these languages. Africans are building that layer themselves.

  • Masakhane, the grassroots pan-African collective, has produced translation benchmarks for more than 30 African languages with no sovereign budget at all.
  • Lelapa AI’s InkubaLM is a 400-million-parameter model trained from scratch for isiZulu, Yoruba, Hausa, Swahili, and isiXhosa, small enough to run without a hyperscaler.

These are the floor: owned in Johannesburg and in community repositories, not licensed from California, and impossible to switch off from abroad.

The economics are moving the same direction. McKinsey projects that by 2030 inference, the running of models, will overtake training to become more than half of all AI compute, and that inference is highly distributable, pushed toward smaller nodes close to the users it serves.

The expensive, centralized layer is training. The layer that scales with every query, and that wants to sit near the people using it, is inference. That is the layer Africa can own.

Ownership Changes Dependence

Owning the model layer does not escape the compute layer. Cassava Technologies is building Nvidia-powered AI factories on the continent, beginning with GPUs in South Africa, which is real and useful and still runs on silicon designed and rationed abroad.

Lelapa ships InkubaLM and still rents its GPUs overseas, working when Silicon Valley sleeps to get the time. The base models, the chips, and the next security patch still arrive from elsewhere.

“Made in Africa” can be a goal, not a law, that changes dependencies, and that matters.

The risk shifts from losing tomorrow what you already depend on, the Fable 5 problem, to missing out on the next frontier upgrade. The first is a kill switch. The second is a slower disadvantage you can plan around.

“Made in Africa” can be a plan to own the floor, the un-revocable layers that decide whether a tool serves Africans, and borrow the frontier for the ceiling on terms you can walk away from. Ownership of the floor does not guarantee benefit. It returns the decision about benefit to Africans.

“Made in Africa” Realities

A single country has no leverage against Nvidia or US export policy. A bloc has some. Smart Africa and the continent’s finance ministers should pool demand for cheap devices and fund a shared commons of open models and local-language data, because twenty ministers negotiating together can set terms one cannot.

Donors should stop financing only application pilots and start funding the floor: open language datasets, shared regional compute, and the small models that run on the devices Africans already own.

And any funder backing an AI health or agriculture tool should be able to answer one question before the money moves: where does this run, who owns the data, and who can turn it off. If the answer is a single foreign vendor, then there is an inherent sustainability risk.

Build the floor on what cannot be switched off from abroad. Borrow the ceiling if you must. Be resilient overall.

The post Borrow the Ceiling While Building Africa’s Artificial Intelligence Floor appeared first on ICTworks.

]]> https://www.ictworks.org/borrow-the-ceiling-while-building-africas-artificial-intelligence-floor/feed/ 0 20456