In Nigeria, a viral clip purported to capture the minister of the Federal Capital Territory threatening to kill electoral officers. Nigerian journalist Rejoice Taddy wanted to investigate the clip circulating on Facebook, but the recording was unclear, and her first attempt to transcribe it failed.

She turned to Dubawa’s AI audio tool, which transcribed the conversation in Nigerian English and Pidgin. Speaking to the Global Investigative Journalism Network, Taddy said it saved her time and allowed her to concentrate on checking the claim.

Following an election campaign involves hours of broadcasts, speeches and interviews, alongside spending records, social media posts and reports from voters. Tools that make this material easier to search and organise could help journalists and civic groups pursue stories they might otherwise struggle to investigate.

Election season brings renewed promises. New candidates set out their ambitions, as incumbents defend their records. Citizens have an opportunity to press their leaders for answers as those seeking office compete for their votes.

Several African initiatives are developing tools that could support that scrutiny. In Nigeria, developers are helping reporters monitor radio broadcasts. In Kenya, research drawing on election reports is exploring automated classification, while a Code for Africa fellowship for Southern Africa is assembling verified information about political figures for use in social media monitoring.

Dubawa’s audio platform grew out of a gap its developers saw in fact-checking. Much of the attention went to claims circulating on social media, with statements made on radio often going unchecked.

The platform can record programmes, transcribe audio and extract claims for investigation. According to the Centre for Journalism Innovation and Development, which runs Dubawa, it supports Nigerian English, Nigerian Pidgin and Ghanaian English. Journalists can use it to review material that would otherwise take hours to transcribe.

That capability can be handy during an election campaign. Journalists and civic actors can easily follow what candidates tell different audiences, revisit disputed remarks and check assertions about their records in office. At the same time, reporters can use it to examine records circulating online.

The work extends across the election cycle. Before elections, journalists and civic organisations must examine campaign conditions, document political intimidation and scrutinise use of public resources. Much of the evidence is often scattered across speeches, radio programmes, official records and reports from communities. As campaigns gather pace, the material can accumulate and exceed the capacity of monitoring teams to examine it.

Afterwards, attention turns to discrepancies in polling-station figures, delays in announcing results, and the handling of legal challenges.Artificial intelligence could help with that workload and African organisations are already exploring and using tools that transcribe broadcasts, organise reports from voters and follow public statements by politicians. These new technologies can widen the reach of election monitoring, particularly for teams with limited staff and large areas to cover.

But on the other hand, watchdogs are also having to confront AI-generated disinformation and other attempts to manipulate public debate. Organisations working to protect civic and political space therefore need to understand the technology both as a source of threats and as a resource for their own work.

Making sense of reports from voters

Kenya’s Uchaguzi initiative illustrates the volume of material involved.

By midnight at the start of election day on 9 August 2022, the platform had collected almost 30,000 reports through social media, SMS, USSD and the web. Staff, volunteers and civil society partners had been working in its election situation room since 1 August. Volunteer teams organised, reviewed and located reports before publishing relevant information.

Researchers later used about 14,000 categorised and geotagged reports from the election to investigate whether language models could assist with that work. The dataset included accounts of alleged official misconduct, counting irregularities and violence. Their paper, published in 2025, examined the use of models to classify reports and identify the places they described.

The research was possible because of the initial painstaking work of documenting, categorising, and organising typse of incidents. Automating some of that preparation could give monitoring teams more time to follow up with the people who submitted reports.

The research builds on years of citizen reporting. Uchaguzi has operated in Kenya since 2010, drawing on contributions from the public and the work of volunteers and partner organisations. Their records now offer material for testing how well language models can handle reports from an actual election.

Following the money and public statements

In Nigeria, Kimpact Development Initiative’s PoliMoney framework combines field observation, media tracking and expenditure estimates to examine how candidates deploy money.

The initiative aims to promote transparency and accountability in campaign spending, including compliance with regulatory limits. Kimpact describes its findings as an overview of spending patterns rather than a financial audit.

Although its published methodology does not identify AI as part of the process, PoliMoney illustrates the detailed observation needed to build a record of campaign expenditure. Researchers must establish what was spent, how they arrived at an estimate and where information is missing.

In South Africa, fellows collected and validated data points and social media identifiers associated with Politically Exposed Persons (PEPs) and Politically Influential Persons (PIPs) through Code for Africa’s 2026 Civic Graph Fellowship, with an emphasis on the nation’s local government elections. They connected platform monitoring and structured civic data using Troll Tracker, a platform driven by natural language processing (NLP) and machine learning (ML).

What organisations need to make it work

These applications are developing as election watchdogs face threats that also involve AI, including fabricated recordings and other misleading political content. Organisations working to protect civic and political space need the capacity to investigate those threats while assessing how the technology might support their own work.

International IDEA’s workshops with electoral bodies and civil society organisations in Malaysia, Albania, South Africa, Panama and Senegal explored responsible AI use, including AI literacy, ethics and human rights, content moderation and electoral management.

Nearly half of participants rated their understanding of AI as low. A third of participating organisations nevertheless reported using it in election-related work. Fewer than one in five respondents said their organisations had internal human rights or ethical review protocols for adopting new technologies.

The findings suggest a need for training and clearer procedures alongside investment in software. Staff need to understand how a tool handles sensitive information and how much checking its outputs will require. For organisations monitoring broadcasts and citizen reports, the accuracy of the tools with local languages will be particularly important.

Funding also has to cover the work around the software. Monitoring teams need trained researchers, access to records and established ways to refer incidents for investigation or response.

AI use in civic monitoring may remain out of reach for many organisations today. The initiatives already under way, however, offer specific approaches that others can test and adapt to their needs. Their usefulness will become clearer over successive campaigns, as teams assess how much time they save and whether they can investigate material that previously went unexamined.

In Nigeria, Dubawa and FactCheckAfrica are deploying AI-enabled verification, including rapid-response WhatsApp claim-checking and audio analysis of live radio broadcasts in Nigerian English and Pidgin.

South Africa’s Media Monitoring Africa runs Real411, a platform where the public reports suspected disinformation and digital harms. During the 2024 election, there was a six-fold increase in complaint volume. Real411 is now folded into a broader “Democracy Shield” partnership with the government.

These initiatives are just a few of those available on the continent. AI provides an opportunity to monitor elections. Beyond biometrics, what can be tracked includes live analysis of polling-station camera/sensor feeds to flag irregularities and analysis of voting-pattern data to detect anomalies.

This data blog was written by Emma Laura Kisa, Project Manager and Data Analyst at Code for Africa’s DataLab, and edited by Kiprotich Koros.

Code for Africa (CfA) is the continent’s largest network of civic technology and data journalism labs, with teams in 21 countries. CfA builds digital democracy solutions that give citizens unfettered access to actionable information that empowers them to make informed decisions and strengthens civic engagement for improved public governance and accountability. This includes building infrastructure like the continent’s largest open data portals at openAFRICA and sourceAFRICA. CfA incubates initiatives as diverse as the africanDRONE network, the PesaCheck fact-checking initiative, the sensors.AFRICA air quality sensor network, and the research and analysis programme CivicSignal.

CfA also manages the African Network of Centres for Investigative Reporting (ANCIR), which provides the continent’s best muckraking newsrooms with the newest possible forensic data tools, digital security, and whistleblower encryption to help to improve their ability to tackle crooked politicians, organised crime, and predatory big business. CfA also runs one of Africa’s largest skills development initiatives for digital journalists, and seed funds cross-border collaboration.

In Senegal, menstruation remains a subject few people discuss openly, and that silence carries a cost for adolescent girls. Civil society organisations are pushing the issue on to the political agenda. But anyone trying to build a case for reform runs into the same wall: the evidence base is thin, dated, and repeated far more often than it is checked.

What the fieldwork actually says

The most substantial work on the subject in Senegal comes from two regional studies commissioned by UN Women and Water Supply and Sanitation Collaborative Council (WSSCC), in Louga in 2014 and in Kédougou in 2015, carried out with support from the ministry of water and sanitation and the ministry of women’s affairs.

The Kédougou report is blunt about infrastructure. Half of the schools it visited had no toilets at all. Only one of the markets surveyed had toilets, and women avoided them because of their condition and because they did not want to share facilities with men. About 40% of the girls interviewed said they missed school at least once a month during their period, and working women described reduced mobility and irregular activity during those days.

The same report frames the scale plainly: between puberty and menopause, women menstruate for roughly 3,000 days.

Two caveats belong with those numbers, and they matter. These are two regions, not a country. And the findings are declarative, what girls and women reported, not what was observed or measured against attendance registers.

The numbers that circulate

One statistic appears in almost everything written about menstruation and schooling in Senegal: that only 1% of schools have separate toilets for girls and boys. It is repeated in NGO project pages and press coverage. The pages that carry it do not cite the survey it comes from, the year it refers to, or the schools it counted.

It may be accurate. The point is that a figure now doing heavy political work has been travelling for years without its paperwork. The same drift affects the 40% absenteeism finding, which was collected in one region in 2015 and is now often quoted as if it described the country.

This is what a thin evidence base does. It does not stop the conversation. It makes the conversation impossible to verify.

There is now a way to measure this properly

Since 2021, the WHO/UNICEF Joint Monitoring Programme (JMP) has maintained harmonised menstrual health indicators in its country database. The framework tracks seven dimensions of menstrual health, from knowledge and access to menstrual materials to privacy, participation, pain management and access to support.

Six of them are built into the seventh round of UNICEF’s Multiple Indicator Cluster Survey, whose menstrual health module launched in 2023. The JMP’s 2025 update on household WASH carries an annex of national estimates covering 70 countries.

Senegal is in that database, but not for this. The JMP’s 2023 progress update lists the 53 countries with nationally representative data on at least one menstrual health indicator between 2014 and 2022. Twenty-two are in sub-Saharan Africa, including Senegal’s neighbours Guinea-Bissau and the Gambia, and nearby Niger, Burkina Faso and Côte d’Ivoire. Senegal is not among them, although it appears elsewhere in the same report for water carriage and drinking water coverage.

Senegal is not short of statistical capacity. It runs a continuous Demographic and Health Survey, the EDS-Continue, managed by the ANSD (Agence Nationale de la Statistique et de la Démographie) and was the first country in Africa to do so.

Nor would it need to design anything. The instrument already exists: the model women’s questionnaire for the eighth phase of the DHS, the template Senegal’s own continuous survey is built on, carries a short menstrual hygiene section, reproduced in full in the JMP’s annex on menstrual health questions. Its core items ask what a woman used during her last period and whether she could wash and change in privacy at home. Adding them is a matter of interview minutes in a sample already being surveyed, not a new data collection exercise.

What a country counts

What a state chooses to count is a statement about what it treats as a public matter. The reform agenda is concrete, free products for schoolgirls, tax relief, renovated school toilets, social workers back in schools, and every one of those measures needs a baseline and a way to check whether it worked.

Until that baseline exists, advocates will keep arguing with figures they cannot source, against budgets defended with figures that are audited. That is not a fair fight, and it is not one they are likely to win.

Advocacy initiatives

Règles sans tabou: Justice Menstruelle Sénégal is an advocacy platform, not an operational NGO. It argues that access to period products is a fundamental right blocked by three things: price, taboo, and missing infrastructure. Its work is informing and campaigning in schools, companies and communities, documenting the policy levers, chiefly the tax question, and covering the cultural history of menstruation. It is recent and still building support.

Worth confirming operational details with them directly, since their site now blocks automated access.

Soyons reglos: (Let’s Talk Period) is a regional awareness campaign covering West and Central Africa, not a Senegal-specific one. It was set up by the Muskoka Fund with WHO Africa, UN Women, UNFPA, UNICEF, UNESCO Dakar and WSSCC. Its work is communication and convening: a digital campaign with influencers, a web series, quizzes and media partners, country-level awareness work such as in Benin, and two events, a conference for International Day of the Girl in October 2020 and a regional symposium in May 2021 whose reports it hosts. It frames menstrual health through UNESCO’s eight factors, which include advocacy and policy.

This data blog was written by Arame Thiam, senior data analyst at Code for Africa’s DataLab, and edited by Kiprotich Koros.

Code for Africa (CfA) is the continent’s largest network of civic technology and data journalism labs, with teams in 21 countries. CfA builds digital democracy solutions that give citizens unfettered access to actionable information that empowers them to make informed decisions and strengthens civic engagement for improved public governance and accountability. This includes building infrastructure like the continent’s largest open data portals at openAFRICA and sourceAFRICA. CfA incubates initiatives as diverse as the africanDRONE network, the PesaCheck fact-checking initiative, the sensors.AFRICA air quality sensor network, and the research and analysis programme CivicSignal.

CfA also manages the African Network of Centres for Investigative Reporting (ANCIR), which provides the continent’s best muckraking newsrooms with the newest possible forensic data tools, digital security, and whistleblower encryption to help to improve their ability to tackle crooked politicians, organised crime, and predatory big business. CfA also runs one of Africa’s largest skills development initiatives for digital journalists, and seed funds cross-border collaboration.

Menstrual Justice in Senegal: What we don’t measure was originally published in Code For Africa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Par Baname Lare

Lorsque j’ai rejoint le programme de bourses AWA Digitalise Youth Project 2026, je pensais avant tout que cette expérience me permettrait d’améliorer mes compétences techniques sur Wikipédia et Wikidata. Au fil des cinq mois du programme, j’ai découvert que la contribution à Wikimédia ne consiste pas seulement à créer des articles ou à atteindre des objectifs chiffrés. Elle consiste surtout à renforcer une communauté, à partager des connaissances et à donner à d’autres les moyens de contribuer à leur tour.

L’un des changements les plus importants dans ma façon de voir le mouvement Wikimédia est venu de mon rôle auprès des autres participants. En plus de produire chaque mois une trentaine d’articles et de créer de nouveaux contenus, j’ai accompagné plusieurs fellows dans l’édition des projets Wikimédia et contribué à former de nouveaux contributeurs. Cette expérience m’a fait comprendre qu’un contributeur ne mesure pas son impact uniquement au nombre d’articles qu’il crée, mais aussi au nombre de personnes qu’il inspire et aide à progresser. Voir un nouveau contributeur publier son premier article ou gagner en autonomie a été tout aussi satisfaisant que mes propres réalisations.

Cette expérience m’a également amené à réfléchir à l’importance des droits numériques en Afrique. En travaillant sur des sujets liés à la cybersécurité, à la protection des données personnelles, à la liberté d’expression, à la désinformation ou encore à l’engagement civique, j’ai pris conscience que ces thèmes restent insuffisamment documentés dans de nombreux pays africains. Les lacunes de contenu ne sont pas uniquement un manque d’informations disponibles ; elles reflètent aussi un manque de contributeurs capables de rechercher, de vérifier et de partager ces connaissances de manière neutre et fiable. Contribuer à combler ces lacunes est donc une façon de soutenir un accès plus équitable au savoir.

Les échanges avec les autres boursiers et avec la Wikimedia Community of Togo User Group ont également marqué mon parcours. J’y ai trouvé des conseils, des encouragements et un véritable esprit de collaboration. Cette solidarité m’a donné davantage de confiance et m’a rappelé que le mouvement Wikimédia repose avant tout sur des personnes qui apprennent les unes des autres. Je ressors de cette expérience avec la conviction que le mentorat et le partage des compétences sont essentiels pour assurer la pérennité de nos communautés.

À l’avenir, je souhaite poursuivre cet engagement en accompagnant davantage de nouveaux contributeurs, en développant des contenus de qualité sur les droits numériques et en continuant à renforcer la participation des communautés francophones africaines aux projets Wikimédia. Cette bourse m’a appris qu’un bon contributeur ne se contente pas d’écrire des articles : il aide aussi d’autres personnes à trouver leur place dans le mouvement. C’est cette leçon que je retiendrai le plus et qui guidera mon engagement dans les années à venir.

Je suis fier d’avoir relevé ce défi et j’espère avoir l’opportunité de contribuer à nouveau à de futures éditions du programme afin de continuer à servir le mouvement Wikimédia et la diffusion des connaissances libres.

Cet article s’inscrit dans le cadre du programme de bourses 2026 de l’African Wikipedian Alliance (AWA), une initiative visant à aider les wikimédiens africains à contribuer à combler les lacunes en matière de connaissances au sein des projets Wikimédia, tout en renforçant leurs compétences, leur expérience et leurs communautés. Dans le cadre de ce programme, les boursiers contribuent aux projets Wikimédia, s’attaquent aux lacunes en matière de connaissances dans divers contextes africains et partagent leurs réflexions sur leurs apprentissages et leurs expériences tout au long du programme.

Le programme de bourses AWA est coordonné par Code for Africa et mis en œuvre en collaboration avec des partenaires qui soutiennent la croissance et le développement des communautés Wikimédia africaines.

By Baname Lare

When I joined the AWA Digitalise Youth Project 2026 fellowship programme, my main expectation was that this experience would enable me to improve my technical skills on Wikipedia and Wikidata. Over the course of the five-month programme, I discovered that contributing to Wikimedia is not just about creating articles or meeting numerical targets. Above all, it is about strengthening a community, sharing knowledge and empowering others to contribute in turn.

One of the most significant changes in the way I view the Wikimedia movement came from my role in supporting the other participants. As well as producing around thirty articles each month and creating new content, I supported several fellows in editing Wikimedia projects and helped to train new contributors. This experience made me realise that a contributor’s impact is not measured solely by the number of articles they create, but also by the number of people they inspire and help to progress. Seeing a new contributor publish their first article or become more independent was just as rewarding as my own achievements.

This experience also led me to reflect on the importance of digital rights in Africa. Whilst working on topics related to cybersecurity, personal data protection, freedom of expression, disinformation and civic engagement, I realised that these issues remain insufficiently documented in many African countries. These content gaps are not merely a lack of available information; they also reflect a lack of contributors capable of researching, verifying and sharing this knowledge in a neutral and reliable manner. Helping to fill these gaps is therefore a way of supporting more equitable access to knowledge.

My interactions with the other fellows and with the Wikimedia Community of Togo User Group also left a lasting impression on my journey. There, I found advice, encouragement and a genuine spirit of collaboration. This solidarity gave me greater confidence and reminded me that the Wikimedia movement is, above all, built on people learning from one another. I come away from this experience convinced that mentoring and the sharing of skills are essential to ensuring the sustainability of our communities.

In future, I hope to continue this commitment by supporting more new contributors, developing high-quality content on digital rights and continuing to strengthen the participation of African Francophone communities in Wikimedia projects. This fellowship has taught me that a good contributor does more than just write articles: they also help others find their place within the movement. This is the lesson I will take away most of all, and it will guide my commitment in the years to come.

I am proud to have risen to this challenge and hope to have the opportunity to contribute again to future editions of the programme so that I can continue to serve the Wikimedia movement and the dissemination of free knowledge.

A French version of this story is available here.

This article is part of the African Wikipedian Alliance (AWA) Fellowship Programme 2026, an initiative supporting African Wikimedians to contribute to closing knowledge gaps across Wikimedia projects while strengthening their skills, experiences, and communities. Through the fellowship, fellows contribute to Wikimedia projects, engage with knowledge gaps across African contexts, and share reflections on their learning and experiences throughout the programme.

The AWA Fellowship Programme is coordinated by Code for Africa and delivered in collaboration with partners supporting the growth and development of African Wikimedia communities.

By Etoe Ruddy Carine Tchuekam

Executive Summary

The current reality of civic engagements is that AI systems are reshaping this advocacy faster than the legal frameworks designed to protect it. Youth human rights defenders face distinct AI-mediated risks; linguistic invisibility, digital surveillance, AI-accelerated disinformation, and financial digitisation, none of which are adequately addressed by existing African human rights or digital governance instruments. This brief argues that closing the governance gap does not require new rights, but the interpretive and procedural adaptation of protections already guaranteed under the African Charter. It proposes five institutional measures and a structured recommendation framework addressed to four distinct actors: the African Commission, African Union and ECOWAS, digital platforms, and HRD organisations themselves.

Note on Methodology

This brief employs a legal gap analysis methodology, examining five regional and international instruments African Charter on Human and Peoples’ Rights Articles 9, 10, and 11; ACHPR Resolution 473 (AI and human rights, 2021); ACHPR Resolution 580 (internet shutdowns); UNESCO 2025 Model Policy Framework for Information Integrity in West Africa and the Sahel; and ICCPR Articles 19 and 22, against the specific conditions of AI-mediated civic engagement in Francophone West Africa and the Sahel. This legal analysis is paired with documented case evidence from Cameroon, Mali, Burkina Faso, and Niger to ground governance gaps in concrete, verifiable failures rather than hypothetical risks. This brief does not propose the implementation of new rights. It identifies where existing rights break down in algorithmic environments and what adaptations are needed to make them functional. The data protection and AI regulation layer, Cameroon Law №2024/017, the ECOWAS Supplementary Act, and the AU Malabo Convention, constitutes a second level of the governance gap examined in the companion analysis developed in July 2026.

1. The Landscape: Youth Civic Engagement Under Algorithmic Pressure

Youth civic engagement in the Sahel is not passive. Movements such as Ñoo Lank in Senegal and anti-Françafrique mobilisations across the region reflect a generation challenging political and economic structures through organised, networked, and increasingly digital forms of participation. WhatsApp groups mobilise community meetings. TikTok documents realtime civic engagements. X (formerly Twitter) coordinates campaigns and public action. Facebook pages build solidarity across borders. For many young people, digital platforms have become the primary infrastructure through which civic participation is organised, amplified, and sustained.

Yet these digital spaces are not neutral public squares. Increasingly, they are governed by algorithmic systems that determine what content is promoted, what information reaches users, and what speech is removed or restricted. Recommendation algorithms shape public discourse, content moderation systems influence visibility, and automated decision-making tools increasingly mediate access to information. At the same time, advances in generative artificial intelligence have lowered the cost of producing manipulated content, synthetic media, and coordinated disinformation campaigns capable of influencing public opinion at unprecedented speed and scale.

A documented example occurred during the 2025 presidential election in Cameroon, where AI-generated images portraying fabricated political crowds circulated widely on social media platforms. These synthetic visuals contributed to the spread of misleading narratives about candidate popularity and highlighted how generative AI can be used to distort political perception in real time during sensitive democratic processes. The incident underscored growing concerns about the ability of existing verification systems and platform moderation mechanisms to respond effectively to AI-generated political content in African electoral contexts.

The consequence is not merely technological but democratic. As civic engagement becomes increasingly dependent on digital infrastructures, the exercise of fundamental rights, including freedom of expression, access to information, freedom of association, and participation in public affairs, is increasingly shaped by systems that operate with limited transparency and accountability.[6] Yet existing regional and national governance frameworks provide little guidance on how these technologies should be designed, deployed, or overseen in ways that protect human rights. The central policy challenge is therefore not how to regulate youth participation itself, but how to ensure that the digital and AI systems shaping that participation remain transparent, accountable, and consistent with democratic values.

2. AI-Mediated Risks to Youth Civic Space: What the Evidence Shows

Exposure 1: Inequality in Digital Visibility and Linguistic Representation

The first structural risk concerns who is visible in digital civic spaces. Africa’s linguistic diversity, with over 2,000 languages, contrasts sharply with the linguistic limitations of most large-scale artificial intelligence systems, which are predominantly trained on high-resource languages such as English and French.

Research in AI governance and computational linguistics confirms that African languages including Mooré, Bambara, Fulfulde, Hausa, and Wolof systematically underperform across AI reasoning, generation, and classification tasks due to their underrepresentation in training datasets. This creates unequal conditions of visibility for civic expression: legitimate civic speech in these languages is misclassified, deprioritised, or removed due to limited linguistic and contextual understanding by content moderation systems, while harmful content may remain insufficiently moderated.

Africa has over 2,000 languages, yet AI moderation systems provide robust support for only a small fraction of Africa’s languages. The result is a structurally uneven information environment in which some voices are amplified while others are rendered effectively invisible in digital civic spaces.

Exposure 2: Digital Surveillance and the Monitoring of Civic Expression

The second risk concerns who is being watched and how civic behaviour is monitored. Across Mali, Niger, Burkina Faso, and Chad, civic space has contracted alongside the expansion of digital monitoring practices targeting civil society actors and online expression.

These practices include social media monitoring, metadata analysis, and the use of automated systems to map communication networks and identify patterns of mobilisation. While not always explicitly defined as artificial intelligence, many of these systems rely on automated data processing and pattern-recognition tools that significantly increase state capacity to monitor civic networks at scale. In contexts where cybersecurity and counterterrorism frameworks are broadly defined, such technologies can heighten risks for youth civic actors engaged in online mobilisation, leading to self-censorship or reduced participation in digital public life.

During the October 2025 presidential election in Cameroon, users reported disruptions in internet connectivity and instability of virtual private networks, raising concerns about evolving forms of network management that affect access to information and communication.

Exposure 3: AI-Accelerated Disinformation and the Integrity of Civic Information

The third risk concerns what information can be trusted in digital civic spaces. Disinformation has become a persistent feature of digital political communication in the Sahel, particularly in contexts of political transition and security instability.

Reports from international monitoring organisations have documented coordinated online influence campaigns in Mali, Burkina Faso, and Niger that amplify political narratives through social media platforms. The emergence of generative AI further increases the scale and sophistication of such content, enabling the creation of realistic synthetic text, audio, and visual materials that are increasingly difficult to distinguish from authentic civic communication. Across Africa and Asia, political campaigners have produced AI deepfakes of foreign leaders endorsing local political parties and candidates, illustrating the cross-border character of AI-assisted disinformation.

For young civic actors who rely on digital platforms for political engagement and mobilisation, these dynamics contribute to an information environment where verification is more difficult, trust is weakened, and civic discourse becomes more vulnerable to manipulation.

Exposure 4: Financial and Administrative Control of Civil Society Sustainability

The fourth risk concerns who can sustain civic engagement over time and under what conditions civil society organisations can continue to operate. Across the Sahel, governments have systematically weaponised financial and registration oversight as tools of civic space restriction, a pattern that becomes substantially more efficient and less visible as administrative and financial systems are digitised and partially automated.

This is not a hypothetical risk. In Mali, the transitional government suspended civil society organisations receiving foreign funding as early as December 2022, initially targeting French-supported NGOs. Since December 2023, the junta has dissolved at least four organisations, including the Observatoire pour les Élections et la Bonne Gouvernance and the Coordination des Mouvements, Associations et Sympathisants de l’Imam Mahmoud Dicko (CMAS), citing public order grounds. By 2025, the junta had suspended all civil society groups receiving foreign funding and introduced draft legislation to tax CSOs, effectively using financial compliance requirements as instruments of suppression.

The governance relevance of this pattern is specific. The legal mechanism in each instance is administrative, registration requirements, funding disclosures, compliance obligations. What digitisation and automation add is efficiency and scale: the capacity to monitor financial flows across a wider range of organisations, flag non-compliance automatically, and trigger suspension or dissolution without the resource-intensive processes that previously limited the reach of such surveillance. Financial digitisation does not create a new form of civic space restriction; it removes the practical limits that previously constrained it.

The governance response required is correspondingly specific: safeguards ensuring that digitised financial oversight mechanisms cannot be deployed against civic actors without judicial authorisation, that automated compliance flags are subject to human review and right of appeal, and that financial monitoring data collected from civil society organisations is not shared with security agencies absent an independent judicial threshold. These protections are not currently provided by any instrument examined in this brief.

In Burkina Faso and Niger, civil society organisations operate under comparable administrative oversight and increasing compliance requirements, while financial systems across the region are becoming more digitised, including electronic banking and automated compliance monitoring tools used within financial institutions.

NOTE: These four exposures are not isolated incidents. They are structurally connected expressions of the same governance failure: legal instruments written before the systems now violating them existed. The question is not whether the risks are real. The focus is what governance architecture can address them, and what existing law already requires, properly interpreted.

3. Governance Pathways: What Existing Law Must Do Now

This brief demonstrates that while existing African and regional human rights instruments provide strong protections for freedom of expression, association, and participation, they were not designed to operate in environments where civic engagement is increasingly mediated by artificial intelligence systems and algorithmic infrastructures. As a result, there is a growing gap between formally guaranteed rights and the practical conditions under which those rights are exercised in digital civic spaces across the Sahel.

Closing this gap does not require the creation of new rights, but rather the adaptation and operationalisation of existing legal protections to reflect contemporary technological realities.

First, the African Commission on Human and Peoples’ Rights should issue an interpretive guidance note clarifying that the rights to freedom of expression and association under Articles 9 and 10 of the African Charter extend to algorithmically mediated environments, including content ranking systems, automated moderation processes, and digital mobilisation platforms.

Second, under Resolution 473 and the emerging African Union AI governance agenda, minimum standards should be developed for any algorithmic system that significantly affects public discourse or civic participation. These standards should require transparency in content moderation and ranking systems, linguistic and contextual adequacy for African languages, and the publication of human rights impact assessments for systems deployed in politically sensitive environments.

Third, Resolution 580 on internet shutdowns should be interpreted and expanded to reflect the evolution of digital interference beyond full network shutdowns. Contemporary restrictions increasingly take the form of targeted throttling, selective platform restrictions, algorithmic downranking, and disruptions to specific communication tools used for civic mobilisation. These forms of interference, while less visible than complete shutdowns, can have comparable effects on civic participation.

Fourth, as financial and administrative systems become increasingly digitised, the African Union and the African Commission should develop safeguards ensuring that digitised financial oversight cannot be deployed against civil society actors without judicial authorisation and is subject to independent review. Automated compliance flags must carry a right of appeal, and financial monitoring data must not be shared with security agencies absent an independent judicial threshold.

Fifth, the African Commission should consider establishing a dedicated Special Rapporteur mandate on digital and AI-mediated civic space, tasked with monitoring emerging risks to freedom of expression, association, and participation arising from algorithmic systems.

Note on the second governance layer: The governance gap identified in this brief operates at two levels. The first, addressed here, is the failure of human rights instruments to govern AI-mediated environments. The second is the failure of data protection and AI regulation frameworks, including Cameroon Law №2024/017, the ECOWAS Supplementary Act, and the AU Malabo Convention, to adequately protect HRDs from AI-mediated surveillance and automated decision-making. That second layer is examined in the companion analysis developed in July 2026.

4. Youth HRDs as Governance Actors: Beyond Risk and Towards Agency

The risks documented above do not exhaust the picture. Across the Sahel, young human rights defenders are not only subject to algorithmic systems, they are actively contesting them, documenting their effects, and building governance alternatives. Organisations such as Africtivistes in Senegal and Burkina Faso have developed local fact-checking and digital literacy infrastructure that directly responds to the disinformation risks described above. Youth-led coalitions in Niger and Mali have produced documentation of content moderation errors affecting their civic communication and used those records to pressure platforms toward greater transparency. In Cameroon, youth activists have developed informal verification networks to counter AI-generated political content circulating during sensitive electoral periods.

This matters for governance design. A framework built only around risk and institutional response will miss the most important resource available for its own implementation: the expertise, networks, and legitimacy of youth HRDs who are already doing governance work on the ground. The recommendations below are therefore not addressed solely to states and regional bodies. They presuppose the active participation of youth HRD organisations in monitoring, documentation, advocacy, and enforcement, as partners in governance, not beneficiaries of it.

5. Recommendations

This brief proposes an adaptive legal framework on AI and civic space grounded in the analytical position advanced throughout: that the Sahel’s existing human rights foundations are substantively adequate, and that what is needed is interpretive and procedural adaptation to govern technologies that did not exist when those frameworks were designed.

The framework is structured around three interdependent legal reforms, with obligations addressed to four distinct actors.

Actor

What They Must Do

African Commission on Human and Peoples’ Rights

Issue an interpretive guidance note extending African Charter Art. 9 and 10 protections to algorithmically mediated environments. Establish a dedicated Special Rapporteur mandate on digital and AI-mediated civic space to monitor emerging risks and report on state and non-state compliance.

AU and ECOWAS

Develop binding minimum standards for algorithmic systems affecting civic participation: linguistic adequacy, transparency in automated decision-making, and mandatory human rights impact assessments before deployment in sensitive contexts. Expand Resolution 580 to cover partial digital interference including throttling, selective platform restrictions, and algorithmic downranking.

Digital Platforms

Publish transparency reports disaggregated by country and language on content moderation decisions affecting civic and political speech. Conduct and disclose human rights impact assessments before deploying or materially modifying algorithmic ranking systems in politically sensitive contexts. Establish accessible appeals mechanisms in relevant national languages. Designate regional civil society contact points for HRD organisations seeking remedy.

HRD Organisations

Develop internal documentation protocols for recording algorithmic interference with civic communication. Build peer monitoring networks across Sahelian countries to aggregate evidence of platform governance failures. Use documented cases to pressure platforms and regional bodies toward greater transparency and accountability. Participate actively in governance consultations as rights-holders, not only as observers.

Reform 1: Interpretive Expansion of Existing Human Rights Standards

African human rights law should be formally interpreted to recognise that the rights to freedom of expression, association, and access to information apply fully to algorithmically mediated environments. This includes recognition that expression includes algorithmic visibility and distribution; association includes digitally networked civic mobilisation; and interference may occur through automated systems that shape access to civic space. This interpretive approach is consistent with the existing jurisprudence of the African Commission, which has progressively extended Charter protections to new contexts without requiring the drafting of new instruments.

Reform 2: Minimum Standards for Algorithmic Systems Affecting Civic Participation

A set of binding minimum standards should govern any artificial intelligence or automated system that materially affects civic engagement. These standards should require linguistic and contextual adequacy in relevant regional languages, transparency regarding automated decision-making processes, documentation of systems affecting civic participation, and human rights impact assessments prior to deployment in sensitive contexts.

Reform 3: Rights-Based Accountability and Remedy Mechanisms

A rights-based governance framework must ensure that individuals and civil society actors have access to effective remedies when AI systems affect their civic participation. This includes the right to meaningful explanations for automated decisions, accessible complaint mechanisms, and independent review procedures. Without such mechanisms, the gap between formally guaranteed rights and their practical exercise in algorithmic environments will continue to widen.

Conclusion

Sahelian states that choose to adopt this framework are not assuming new obligations, they are operationalising commitments already made under the African Charter, adapted to reflect the technological conditions under which those commitments must now be kept. Without such adaptation, formally guaranteed rights will become progressively harder to exercise in practice within algorithmically mediated civic spaces. The governance response at the data protection and AI regulation level, and the practical tool that enables every actor in this governance chain to assess AI systems before deployment, is developed in the companion analyses that follow.

Beyond Surveillance: Building Rights-Based AI Governance for Youth Civic Engagement in the Sahel was originally published in Code For Africa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Before Nakuru began building a denser air-quality monitoring network, a complaint about pollution could be true, urgent and still difficult to act on.

Residents could point to smoke from burning waste, emission from traffic, dust rising in the wake of dozens of lorries ferrying sand through unpaved roads, emissions from traffic or the smell drifting from an industrial area. Health workers could speak about the burden of respiratory diseases from residents and county officers could inspect a site or ask companies for reports. But without a continuous stream of local data, it was hard to enforce existing standards and regulations, show how pollution changed from one neighbourhood to another, which areas were repeatedly affected, or whether a problem was an isolated incident or part of a pattern.

“Before, we didn’t have evidence,” said Simon Muthami, an environment officer with the Nakuru County Government. “You could say you are producing air pollution. But now with the data you are able to show that this place is highly polluted compared with this one.”

In 2025, sensors.AFRICA installed 22 low-cost air-quality sensors across the city as part of RESPIRA, a project that combines sensor readings, community reports, meteorological data and machine-learning tools to understand what residents are breathing and how public institutions might respond.

The project is a collaboration between Egerton University, KU Leuven, Nakuru County Government, Kenya Meteorological Services Authority, sensors.AFRICA and other partners. The project is trying to answer the question whether a growing city can turn air-quality data into action before pollution becomes normalised as the cost of urbanisation.

Nakuru was chosen because it represents a problem many African cities are facing with far less attention than capitals receive. It is growing rapidly, sits on a major transport route between Nairobi and western Kenya, and has all the hallmarks of urbanisation: increased traffic, construction, waste, settlements, industrial activity and a greater need for public services.

“Nakuru is a rapidly developing secondary city,” said Dr Nick Rahier, an environmental anthropologist at KU Leuven and one of RESPIRA’s principal investigators. “Compared to Nairobi, there was only limited focus on air quality.”

For years, Nakuru’s understanding of air pollution rested on piecemeal evidence. Data was gathered in specific studies, equipment was trialled in specific sites, and residents made observations about what they could see, smell or feel in their neighbourhoods. Much of that information was useful, but it was rarely enough to give the city a continuous picture of how air quality changed across the city, or how those changes could guide planning and enforcement.

RESPIRA is attempting to build something more lasting. Its sensors measure particulate matter, including PM2.5 and PM10, with the project placing particular emphasis on PM2.5, the fine particles small enough to lodge deep in the lungs. The data is combined with weather information and residents’ reports submitted through AngaWatch, a citizen-reporting tool used by ward-based climate ambassadors.

But even the expanded footprint of over 22 air quality sensors is not enough to capture the full complexity of Nakuru County. The existing network is still concentrated mainly in and around the city, while other parts of the county have different pollution sources, settlement patterns, transport pressures and economic activities that remain less understood. Expanding the network would give people a fuller picture of how air quality varies beyond the urban core, but it would also require more partners, stronger county support and sustained funding.

For Alicia Olago, senior product manager at sensors.AFRICA, the question is not whether the model can be expanded, but whether the resources exist to sustain that expansion.

“Nakuru is huge,” she said. “Of course it is possible to scale, but the biggest thing is that we need more partners, the support of the county government and funding.”

Low-cost sensors make wider monitoring feasible, but they are not a silver bullet. Apart from the initial cost, they also require calibration, maintenance and careful interpretation. They also depend on the infrastructure around them. Olago said the devices can transmit data through Wi-Fi or cellular networks, but Wi-Fi is not available in all the places where monitoring is needed. The project therefore uses SIM cards, including IoT SIMs, although mobile connectivity can also be unreliable in some locations. When that happens, the sensors store readings on SD cards and transmit them when the network is restored.

There are other challenges too. Prof Wilkister Nyaora Moturi of Egerton University, RESPIRA’s principal investigator, said vandalism has forced the project team to look for secure places to install sensors.

The data gathered shows a mixed picture of Nakuru’s air quality.

“The quality of air ranges from moderate to good,” Moturi said, adding that readings can become poor in some locations and at certain times. “Air quality is not constant. It is affected by weather, it is affected by the activities of the day.”

In dry periods, dust can push particulate levels higher. During peak traffic, busy roads may show worse readings. In residential areas, waste burning and cooking fuels contribute to exposure. Garages, roadside activity, quarrying routes and industrial zones produce other patterns. A citywide average can make the situation appear manageable, but residents breathe in specific places and at specific times.

Moturi warned that even moderate air quality should not be treated as harmless.

“It doesn’t mean that we should be happy with moderate,” she said. “Low levels of pollutants over a long period of time can lead to chronic issues.”

That is why RESPIRA is not relying on sensors alone. The project’s climate ambassadors collect accounts from residents in different wards, linking technical readings to lived experience. Many of the ambassadors are connected to existing Ward Climate Change Committees, which gives the project a route into local governance structures rather than treating residents as people who merely supply anecdotes.

Rahier described the approach as a combination of sensor readings and citizen narratives.

“We have already collected a lot of data, both sensor readings about the quality of the air, and stories about people’s experiences with polluted air,” he said.

That combination was at the heart of the RESPIRA Nakuru Sense-Making Workshop, held from 24 to 27 August 2026. Researchers, county officials, meteorologists, health practitioners, urban planners and community ambassadors gathered to examine sensor data, residents’ reports and technical analysis. The meeting was not meant simply to review what had been collected, Rahier said, but to ask what the city could now do with it.

“We are also thinking about what now we can do with this data,” he said. “Do we move to policy?”

County officials say that having the data will help them move closer to action. Muthami said Nakuru already has environmental rules it can use, including national environmental law and county legislation on waste management and climate change. Companies can be required to submit emissions reports, and county officers can recommend corrective measures where standards are not met. But laws depend on implementation, and implementation depends partly on evidence.

“The policies are there, the laws are there,” Muthami said. “Just for the government now to be very strict on enforcement.”

Grace Karanja, Nakuru County’s Director for Environment and Climate Change, said the county’s air-quality work did not begin with RESPIRA. Earlier collaborations had helped install a smaller number of sensors in strategic locations, including high-traffic areas and public facilities. RESPIRA expanded that foundation and brought in community ambassadors to help identify additional monitoring sites.

“We can now make better decisions because we already have the data,” Karanja said. “Before, it was by word of mouth.”

For Karanja, the data can help the county advise residents, identify areas where air quality is worsening, plan mitigation measures and connect environmental monitoring to the constitutional right to a clean and healthy environment. But the current network still covers mainly the city, while Nakuru County has 11 sub-counties and several other centres where industrial activity, transport pressure and settlement growth are increasing.

“Our vision is to monitor air quality in the whole county,” she said.

Indoor air pollution remains another gap. Outdoor sensors can show what is happening along roads, near public facilities and across neighbourhoods, but household exposure from cooking fuels remains harder to capture. Karanja and other officials note that some of the people most exposed to indoor smoke, including women and children, may not be reached by outdoor monitoring alone.

The next step for RESPIRA is to move from measuring air quality to forecasting it.

Joshua Musau, the project’s data scientist, said the team is using sensor readings, meteorological data and machine-learning models to predict air quality up to 24 hours ahead. If the system is adopted by meteorological authorities, air-quality information could eventually sit beside ordinary weather updates.

“For a long while, AI and machine learning has remained a buzzword,” Musau said. “But what we’re doing in Nakuru is a case to watch.”

A daily air-quality forecast could, for example, help people with asthma decide when to travel or exercise, guide athletes choosing training routes, inform schools about outdoor activity and give health workers another layer of information when tracking respiratory complaints.

“Normally you get alerts on rainfall,” Musau said. “You can have an extra detail of air quality.”

Dr Richard Mwita, Deputy Director of Research and Education of the Kenya Meteorological Services Authority sees this as part of a larger ambition. Kenya has monitored greenhouse gases, studied short-lived climate pollutants, piloted particulate-matter sensors and carried out indoor pollution assessments, he said. But air-quality data remains scattered across research projects, institutions, student studies and temporary pilots.

Dr Mwita says this fragmentation is not unique to Nakuru. In Kenya, air-quality evidence has often been produced through individual studies, short-term projects and institutional pilots, each adding useful information without necessarily becoming part of a shared national system. Without a common framework, a maintenance budget and institutional ownership, the data can remain isolated while the equipment eventually falls out of use.

Dr Mwita said acquiring sensors is often easier than sustaining them. A credible system needs calibration, repairs, field inspection, data storage, modelling, public communication and money set aside after the project period ends.

“The impact of air pollution is not instant,” he said. “It is long term.”

He described air pollution as a “silent killer” because its effects accumulate gradually and rarely create the immediate alarm associated with floods, storms or road accidents. That is why data becomes critical. Where pollution is not measured, it is easier to dismiss and communities struggle to make demands where data is fragmented.

Mwita said scientists and public institutions should not avoid uncomfortable evidence because it may affect reputations, property markets or political interests.

“The facts remain facts,” he said. “We have a problem that is actually a silent killer.”

Nakuru’s new evidence base on air pollution gives officials and residents more than one way to understand the problem. Sensor readings can identify where particulate levels are rising, community reports can help explain what is happening on the ground, and forecasts could eventually warn people before conditions deteriorate. But the same evidence will only become useful if it is taken up by the institutions that can act on it.

Still, there are still many challenges to surmount. RESPIRA may be strengthening Nakuru’s ability to understand its air, but the system is still some distance from the countywide service officials say they want to build. The current network does not yet cover all parts of Nakuru County, and data on household exposure from cooking fuels is still limited. The flow of data also depends on reliable connectivity, secure installation sites and regular maintenance. Even where the evidence is strong, its value will depend on whether county agencies and regulators are willing to use it in planning, inspections, budgets and enforcement when readings point to specific sources or persistent hotspots.

This blog was written by Kiprotich Koros, copy editor at Code for Africa.

Code for Africa (CfA) is the continent’s largest network of civic technology and data journalism labs, with teams in 21 countries. CfA builds digital democracy solutions that give citizens unfettered access to actionable information that empowers them to make informed decisions and strengthens civic engagement for improved public governance and accountability. This includes building infrastructure like the continent’s largest open data portals at openAFRICA and sourceAFRICA. CfA incubates initiatives as diverse as the africanDRONE network, the PesaCheck fact-checking initiative, the sensors.AFRICA air quality sensor network, and the research and analysis programme CivicSignal.

CfA also manages the African Network of Centres for Investigative Reporting (ANCIR), which provides the continent’s best muckraking newsrooms with the newest possible forensic data tools, digital security, and whistleblower encryption to help to improve their ability to tackle crooked politicians, organised crime, and predatory big business. CfA also runs one of Africa’s largest skills development initiatives for digital journalists, and seed funds cross-border collaboration.

Nakuru can now measure its air quality, but the harder part is what comes next was originally published in Code For Africa on Medium, where people are continuing the conversation by highlighting and responding to this story.

I began the AWA EPD-DDI Enhancing Youth Voices and Youth Civic Engagement in a Secured Digital Space fellowship believing I already understood the terrain.

I had worked within Wikimedia communities, led teams and supported projects beyond Nigeria. I expected the fellowship to strengthen my technical skills in finding notable topics, identifying red links and helping contributors develop strong Wikipedia articles.

Instead, its most important lesson had little to do with editing tools. It taught me how knowledge moves across Africa — and how language barriers, limited access to sources and unequal documentation practices can disrupt that movement.

The gap within the gap

My perspective began to change while working on topics connected to countries including Cameroon, Ethiopia and Sudan.

The problem was not always a lack of information. In many cases, valuable knowledge existed, but the sources were difficult to access, verify across languages or adapt to the standards required by English Wikipedia.

This challenged one of my long-held assumptions that content gaps were primarily caused by missing knowledge. During the fellowship, I came to understand that these gaps are also shaped by language, discoverability and the limitations of the tools we use to make knowledge accessible.

This realisation changed how I understand Wikimedia’s knowledge gaps. When a topic from an underrepresented community is missing, it does not necessarily mean contributors are unwilling to document it or that the public lacks interest. Sometimes the knowledge exists, but the path to verifying and presenting it is considerably harder.

This raises a few questions for communities documenting African subjects. Wikimedia’s mission invites us to bring more of the world’s knowledge into view. Yet contributors often encounter policies designed for contexts with extensive, easily searchable records, not places where information is scattered across languages, institutions and formats.

Knowledge equity, therefore, requires us to examine the conditions that determine whose knowledge can be easily found, verified and shared.

Remote learning has its limits

I also realised during the fellowship that mentoring new contributors online is harder than I had acknowledged.

Being in physical community spaces means that I can often recognise confusion, hesitation or excitement as it happens. I can adjust my approach, answer questions and offer reassurance. Online, it is difficult to know when someone is lost, distracted or discouraged.

This experience forced me to reconsider both my patience and my teaching style. I now understand that effective remote mentorship requires intentional feedback, well-paced learning and stronger support systems. It is not enough to transfer an in-person programme onto a digital platform and expect the same results.

Supporting contributors to persist

I leave the fellowship with a sharper understanding of Wikimedia’s knowledge gaps and a more grounded sense of what meaningful contribution demands.

Pan-African initiatives cannot assume that every context is equally searchable, documented or legible to global verification systems. If we are serious about knowledge equity, we need better tools, greater sensitivity to local contexts and communities that are supported not only to contribute, but also to persist when the work becomes difficult.

Looking ahead, I will approach multilingual projects and remote programmes differently. I am most excited by the possibility of a future in which African contributors are equipped to navigate and challenge unfavourable systems, rather than being pushed away by them and leaving knowledge gaps to grow.

The fellowship reminded me that closing Africa’s knowledge gaps is not simply a matter of adding more articles. It is also about building fairer pathways through which African knowledge can be found, verified and shared with the world.

Code for Africa (CfA) is the continent’s largest network of civic technology and data journalism labs, with teams in 21 countries. CfA builds digital democracy solutions that give citizens unfettered access to actionable information that empowers them to make informed decisions and strengthens civic engagement for improved public governance and accountability. This includes building infrastructure like the continent’s largest open data portals at openAFRICA and sourceAFRICA. CfA incubates initiatives as diverse as the africanDRONE network, the PesaCheck fact-checking initiative, the sensors.AFRICA air quality sensor network, and the research and analysis programme CivicSignal.

CfA also manages the African Network of Centres for Investigative Reporting (ANCIR), which provides the continent’s best muckraking newsrooms with the newest possible forensic data tools, digital security, and whistleblower encryption to help to improve their ability to tackle crooked politicians, organised crime, and predatory big business. CfA also runs one of Africa’s largest skills development initiatives for digital journalists, and seed funds cross-border collaboration.

As Ebola spreads through conflict-affected communities, local data is helping reveal where cases are rising, how the outbreak is changing and where the response is under greatest strain.

Before health authorities could confirm the latest Ebola outbreak in the Democratic Republic of the Congo (DRC), samples collected in the country’s east had to be transported to a national laboratory in Kinshasa.

The journey delayed confirmation of what was making people ill. It also delayed the information health workers needed to isolate patients, trace their contacts and determine where the disease might spread next.

Testing has since been decentralised, with laboratories in and around affected communities able to process more samples and return results more quickly.

But faster testing is only part of the information challenge. Data about the outbreak is still spread across situation reports, public datasets and updates from different organisations. This can make it difficult to follow how cases, deaths and recoveries are changing across locations and over time,

To make that information easier to find and explore, openAFRICA has brought together relevant Ebola data in an interactive tracker.

The dashboard brings together figures on confirmed cases, deaths, recoveries, affected health zones and other indicators. It allows users to move beyond a single national total and examine how the outbreak is evolving across different locations.

That local picture matters in an emergency that has grown rapidly since it was declared in May 2026.

By the end of June, DRC had recorded 1,333 confirmed cases and 399 deaths, a case fatality ratio of 29.9%. The figures represented a steep rise from 17 June, when WHO had reported 896 confirmed cases and 232 deaths in DRC. Part of that increase resulted from expanded testing capacity, which allowed laboratories to process a backlog of samples collected earlier in the outbreak.

Health workers face an especially severe risk. Their jobs bring them into close contact with people who may be infectious, sometimes before Ebola has been diagnosed and in facilities where protective equipment, isolation space or trained personnel may be limited.

The Democratic Republic of Congo (DRC) has confronted Ebola 17 times since the virus was first identified in the country in 1976. That experience has helped it develop considerable expertise in testing, surveillance and community engagement. Yet each outbreak presents a different set of challenges.

The current outbreak declared in May 2026 is caused by the relatively rare Bundibugyo virus, for which there is currently no approved vaccine or specific treatment. It is unfolding amid armed violence, mass displacement, strained health services and high levels of movement within DRC and across its borders. These conditions make it harder to identify infections, trace contacts and reach communities with reliable information.

The outbreak began in Ituri but has since spread into North Kivu and South Kivu, as well as across the border into Uganda.

The openAFRICA tracker allows users to examine the outbreak through several connected datasets including confirmed cases, affected health zones, population figures, health facilities and patterns of movement around key regions.

This can help journalists, researchers and public health organisations ask more useful questions: Where are cases rising most rapidly? Which health zones are reporting the most deaths? Where might population movement increase the risk of further spread? Which areas may need closer scrutiny because reported figures remain unusually low?

The data will not provide the complete picture because reporting may be delayed, testing capacity varies and insecurity can prevent surveillance teams from reaching some communities. This data should therefore be understood as the best available account of the outbreak rather than a record of every infection.

Find relevant Ebola data here. The data collated shows cases, mobility index and movements around key regions, health facilities, health zones, and population numbers. From different players. Case data is updated bi-weekly.

Click below to view the interactive dashboard for up-to-date data about the public health emergency.

DRC · Uganda - Ebola Bundibugyo 2026 · Interactive Dashboard

Code for Africa (CfA) is the continent’s largest network of civic technology and data journalism labs, with teams in 21 countries. CfA builds digital democracy solutions that give citizens unfettered access to actionable information that empowers them to make informed decisions and strengthens civic engagement for improved public governance and accountability. This includes building infrastructure like the continent’s largest open data portals at openAFRICA and sourceAFRICA. CfA incubates initiatives as diverse as the africanDRONE network, the PesaCheck fact-checking initiative, the sensors.AFRICA air quality sensor network, and the research and analysis programme CivicSignal.

CfA also manages the African Network of Centres for Investigative Reporting (ANCIR), which provides the continent’s best muckraking newsrooms with the newest possible forensic data tools, digital security, and whistleblower encryption to help to improve their ability to tackle crooked politicians, organised crime, and predatory big business. CfA also runs one of Africa’s largest skills development initiatives for digital journalists, and seed funds cross-border collaboration.

Mapping DRC’s Ebola outbreak was originally published in Code For Africa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Opening a Wikipedia account is easy. Learning to use it well is not.

A new contributor must learn how to judge the reliability of a source, navigate unfamiliar editing tools and make changes without disrupting the work of others. This often brings up harder questions. How should an editor respond to artificial intelligence that produces polished prose and invented citations with equal confidence? How can a patroller protect Wikipedia from vandalism without driving away a beginner who has made an honest mistake? What does it take for an African language to become not merely visible online, but usable across the systems on which digital knowledge increasingly depends?

These questions are at the heart of the work of the African Wikipedian Alliance (AWA): building communities capable of creating, protecting and sustaining their own knowledge.

Between February and May 2026, the African Wikipedian Alliance brought together 256 participants from its Anglophone and Francophone communities for eight learning sessions. The subjects ranged from AI and source editing to gender inclusion, storytelling, Wikidata and African-language infrastructure.

For many people, Wikipedia is the first stop when they want quick information about a subject, event or public figure. Far fewer know about AWA, a community of contributors, organisers, and knowledge advocates working to strengthen participation, leadership, and knowledge equity across Africa’s Wikimedia ecosystem.

When producing text becomes easier, judgement matters more

The arrival of generative AI has unsettled some of the basic assumptions of online knowledge production. It can translate a passage, improve its grammar or suggest the structure of an article in seconds. But, as we have often seen, it can also fabricate a source, distort a fact or produce a confident answer that is not supported by evidence.

For Wikipedia, a platform built on verifiability, this presents both an opportunity and a problem..

During AWA’s February anglophone session, “AI + Wikipedia: The Good, the Bad, and the Best Practices,” participants examined how AI tools can assist with translation, summarisation, grammar and the early structuring of articles. They also considered the risks: fabricated facts, invented citations and misleading content that may initially appear credible.

Rather than presenting AI as either “good” or “bad,” the discussion encouraged contributors to think critically and ethically about how these tools are used. AI can assist editors, but it cannot replace human judgment, verification, and responsibility. At a time when misinformation and synthetic content continue proliferating online, the lesson was that AI is only useful when people know how to use it responsibly.

The francophone community tackled a related subject of patrolling. Patrollers review recent changes, detect vandalism and help maintain the quality of Wikipedia’s articles. At their best, they are caretakers of collaborative knowledge: people who protect quality while helping new contributors understand where they went wrong.

A damaging edit is not always an act of vandalism. It may come from a newcomer who does not yet understand Wikipedia’s rules or technical conventions. Treating those situations in the same way can protect a page at the expense of someone who was trying to improve it.

In both sessions, the technical tools were secondary. The real subject was responsibility: when not to trust a machine, when to undo a contribution and when a new editor simply needs guidance. Editorial quality and community health are not competing priorities. Over time, each depends on the other.

A language needs more than articles

If the February sessions were concerned with protecting knowledge, the language sessions asked who has the infrastructure to produce it in the first place.

Much of the internet treats language as content: words that can be translated and placed on a page. But for a language community seeking a sustained presence on Wikimedia, translation is only one layer of the work.

Editors need an interface they can use. They need agreed terminology, reliable sources and a community large and active enough to sustain a project. The language must also be represented in digital systems that can recognise its words, grammatical forms and meanings.

Expanding African languages and knowledge representation

AWA’s March sessions introduced participants to the Wikimedia Incubator and TranslateWiki. The Incubator allows communities to begin developing Wikimedia projects in languages that do not yet have independent sites. TranslateWiki enables contributors to localise the interfaces through which people navigate and edit those projects.

The work can appear administrative: translating menus, testing pages and building a consistent body of content. Yet these are the foundations on which an independent language project may eventually stand.

AWA returned to the question in April through a session on lexicographical data in Wikidata. Participants were introduced to lexemes, forms and senses — the structures used to describe a word, the shapes it takes and the meanings it carries in a machine-readable form.

Here is why the distinction matters: A word recorded only in an article can be read by the person who finds it. A word represented as structured data can potentially be connected to dictionaries, translation systems and other language technologies. It becomes easier to search, reuse and link across platforms.

Language work in this sense, is infrastructure work. It determines which communities can do more than appear on the internet: which can participate in building the systems through which knowledge will be organised and retrieved.

For African languages that remain poorly represented online, the stakes extend beyond Wikimedia. As digital tools become more deeply embedded in education, communication and public life, languages absent from their underlying systems risk becoming less usable in the spaces those tools create.

Being admitted is not the same as belonging

Building the right infrastructure also means building strong communities and creating the social conditions that allow people to participate consistently.

During AWA’s International Women’s Day discussions in March, contributors, fellows, organisers and movement leaders considered what sustainable participation looks like for African women in digital knowledge communities.

Training was part of the answer, but only part. A person may learn how to contribute and still lack the time, financial support, mentorship or recognition needed to remain involved. She may be welcomed as a participant without being given a meaningful opportunity to lead.

The discussions identified access, visibility, funding and mentorship as interconnected needs. They also raised the importance of care work: the often-unrecognised labour that supports families, organisations and communities, but can limit the time women have available for volunteer-led digital projects.

Inclusion cannot be measured simply by whether the door has been opened. A new editor is not integrated into a community simply because an account has been created. They need the resources, relationships and room to grow.

Communities must be able to tell their own stories

Some forms of community work can be easily counted: articles edited, references added or data entries created. Other forms such as mentorship or organising local networks are however invisible. Processes that are developed by one group can often disappear because the experience was not documented.

AWA’s April Francophone session focused on Diff, the Wikimedia movement’s platform for community news and learning. Participants considered how to structure an article, record measurable results and use photographs and testimony to communicate what a project accomplished.

When community work goes undocumented, the lessons can disappear, organisers cannot learn from the experience and future participants cannot see the history they are joining.

Storytelling is therefore part of a community’s infrastructure, too. It preserves memory, circulates knowledge and allows people to describe their work in their own terms.

That capacity is especially important in a movement concerned with representation. Communities seeking greater control over how African knowledge appears online also need greater control over how their own contribution to that work is understood.

From using tools to shaping a movement

By May, the programme had turned towards the mechanics of independent contribution.

To an experienced editor, headings, references, templates and wiki markup are just ordinary elements of a Wikipedia page. But to a newcomer looking at dense source code for the first time, they can make the platform feel closed and forbidding.

In the Anglophone source-editing session, participants were shown how breaking a page into its component parts can make an apparently complicated whole become a sequence of manageable decisions.

The francophone Wikidata session approached independence from another direction. Participants learned how structured data connects information across Wikipedia, Wikimedia Commons, Wikivoyage and other projects. A well-structured item can help knowledge travel across languages and platforms rather than remaining confined to a single page.

These are practical skills that every contributor should have, but their knock on effects are much greater. Contributors who understand the tools can work with greater independence, and, with experience, they can help newcomers, lead training sessions, organise projects and strengthen the systems on which others rely.

One session is rarely enough, and even eight sessions cannot resolve the structural barriers facing African Wikimedia communities. Training is just a starting point — one that can open a path towards transformation.

Monthly meetups can help people move from observing the Wikimedia movement to contributing to it — and, in time, to shaping where it goes next.

Join the AWA community here to connect with Wikimedians across Africa and take part in future learning sessions.

Code for Africa (CfA) is the continent’s largest network of civic technology and data journalism labs, with teams in 21 countries. CfA builds digital democracy solutions that give citizens unfettered access to actionable information that empowers them to make informed decisions and strengthens civic engagement for improved public governance and accountability. This includes building infrastructure like the continent’s largest open data portals at openAFRICA and sourceAFRICA. CfA incubates initiatives as diverse as the africanDRONE network, the PesaCheck fact-checking initiative, the sensors.AFRICA air quality sensor network, and the research and analysis programme CivicSignal.

CfA also manages the African Network of Centres for Investigative Reporting (ANCIR), which provides the continent’s best muckraking newsrooms with the newest possible forensic data tools, digital security, and whistleblower encryption to help to improve their ability to tackle crooked politicians, organised crime, and predatory big business. CfA also runs one of Africa’s largest skills development initiatives for digital journalists, and seed funds cross-border collaboration.

What it takes to build a Wikimedia community was originally published in Code For Africa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Across parts of Africa, some of the data needed to understand women’s rights, civic participation and democratic governance is still fragmented, hard to access or missing altogether. It sits in offline archives, scattered reports, incomplete records, or systems that are difficult for activists, journalists, researchers and civic actors to use.

The WanaData Fellowship was created to help close that gap.

Through the fellowship, women data scientists and data storytellers are turning civic and gender-related information into structured datasets, evidence-led stories and digital resources that can support advocacy, research, fact-checking, civic technology and public debate.

The first cohort has already shown what this can look like in practice. Fourteen fellows from the Sahel, West Africa and the Horn of Africa produced 80 data-driven stories on issues including digital rights, gender-based violence, climate activism, sexual and reproductive health, humanitarian displacement and democratic governance.

For women activists, journalists, human rights defenders, content creators and civic actors, this work matters because visibility is often the first step towards accountability. When data is missing, fragmented or inaccessible, the issues affecting communities become easier to ignore and harder to challenge.

The WanaData Fellowship is part of the Digital Democracy Initiative (DDI), led by Code for Africa (CfA), in partnership with the European Partnership for Democracy (EPD), AfricTivistes, CFI Media Development, the World Scout Bureau Africa Regional Office and the Kofi Annan Foundation, with contributions from the Netherlands Institute for Multiparty Democracy (NIMD). The initiative works to strengthen and protect democratic spaces and human rights in the digital age.

A fellowship designed for civic impact

The WanaData Fellowship focuses on digital democracy, civic participation, gender and technology, freedom of expression, human rights, digital mobilisation, fact-checking, civic technology innovation, data protection, and Sustainable Development Goal 16 on peace, justice and strong institutions.

It specifically targets regions where access to structured civic and gender-related data remains limited, including countries in the Sahel, West Africa and the Horn of Africa.

Over two years, the programme will support 28 fellows across two cohorts. Each cohort runs for six months and brings together 14 participants:

• Seven data scientists specialising in data structuring, analysis and AI training
• Seven data storytellers specialising in advocacy, digital campaigns and civic technology development

Participants are selected based on merit and country representation.

The programme invests in Africa’s democratic resilience through training, mentorship, peer learning and financial support. Fellows gain practical tools to amplify civic voices and strengthen digital advocacy in their communities.

Turning scattered information into evidence

In many communities, important civic and gender-related information still sits in offline archives, scattered documents or systems that are difficult to access and use. Through the WanaData Fellowship, participants help turn this information into structured digital resources that can support advocacy, research, innovation and informed decision-making.

By digitising archives and organising civic data, fellows contribute to stronger fact-checking ecosystems, better access to reliable information, and the development of civic technology and AI-supported solutions.

This work is especially meaningful for women activists, human rights defenders, content creators and informal civic actors whose voices are often underrepresented in digital spaces. Apart from amplifying these voices, the fellowships also equips participants with the digital and data skills needed to create lasting impact within their communities.

A hands-on fellowship experience

The WanaData Fellowship is designed to be practical, collaborative and impact-driven from the start.

Throughout the six-month programme, fellows work closely with mentors, data experts and fellow participants while gaining hands-on experience in civic technology and digital advocacy.

The journey begins with orientation sessions and peer mentorship, where participants receive technical guidance and connect with a network of changemakers across the continent.

Fellows then move into gender data mapping, identifying gaps in civic and gender-related information within their local contexts and exploring how better data can support more inclusive governance.

The programme also encourages collaboration with institutions and data producers to promote stronger standards, responsible data practices and open access to information.

One of the fellowship’s practical components is the Data Liberation scraperthons, where participants work together to digitise offline archives and transform them into structured datasets that can support advocacy campaigns, civic innovation and research.

By the end of the programme, the data produced through the fellowship becomes a resource for researchers, technologists, journalists, activists and decision-makers working to strengthen evidence-based civic action across Africa.

What the first cohort produced

Fourteen fellows from across the target regions were part of the first cohort, including 12 francophone participants and two anglophone participants. They came from countries across the Sahel, West Africa and the Horn of Africa, including Senegal, Mauritania, Mali, Burkina Faso, Niger, Chad, Sudan, Guinea, Benin, Togo, Cameroon, South Sudan, Ethiopia and Somalia.

Together, they produced 80 data-driven stories on digital rights, gender, civic participation and democratic governance.

Some stories examined the cost of gender-blind data systems and how missing or incomplete data can obscure violence and harm women’s rights. Others explored how digital communities are reshaping climate activism in Africa, how online gender-based violence affects women’s safety, how misinformation influences access to sexual and reproductive health services in Burkina Faso, and how women and girls are affected by Mali’s humanitarian crisis, where 78% of displaced people are women.

These stories help make underreported issues more visible, giving activists, journalists, researchers and communities evidence they can use in public debate, advocacy and accountability work.

Through cross-country collaborations, fellows also showed how local knowledge, data skills and storytelling can come together to widen participation in democratic spaces.

Why this work matters

Digital spaces are now central to how people engage with democracy, access information and advocate for change. But women-led civic movements still face barriers to visibility, reliable data, digital inclusion and meaningful participation.

The DDI WanaData Fellowship responds to this gap by bringing together data science, storytelling, civic technology and advocacy.

It gives emerging changemakers the opportunity to build practical skills while contributing to a wider movement for inclusive digital governance across Africa. For aspiring data scientists, digital advocates, technologists and civic innovators, the fellowship is a chance to collaborate across borders, build meaningful solutions and help shape the future of democratic participation in the digital age.

Join the movement

The WanaData Fellowship shows what becomes possible when collaboration, civic engagement and technology are rooted in local knowledge.

At Code for Africa, we believe local talent and community-driven innovation are essential to building more transparent, inclusive and resilient societies. Through initiatives such as the Digital Democracy Initiative, a new generation of civic innovators is already using data and digital tools to make overlooked issues more visible, strengthen advocacy and expand participation in Africa’s democratic spaces.

This data blog was written by Arame Thiam, senior data analyst at Code for Africa’s DataLab, and edited by Kiprotich Koros.

How women are using data to strengthen Africa’s digital democracy was originally published in WanaData Africa on Medium, where people are continuing the conversation by highlighting and responding to this story.