25 Aug 2026 Weaving the digital web: Africa’s sovereign AI, community networks, and the ethos of Ubuntu
Rebecca Ryakitimbo
The African continent stands at a technological crossroads. As the global race for Artificial Intelligence dominance intensifies, Africa is moving beyond the role of a passive consumer of algorithms. Instead, a continent-wide movement is emerging, one that seeks to reclaim sovereignty over data, language, and innovation. At the heart of this transformation is a fundamental realignment of technology with the foundational African values of Ubuntu (interconnectedness), Ujamaa (familyhood/cooperation), Harambee (pulling together), and the Kwanza principles of Umoja (unity) and Kujichagulia (self-determination).1
C
onnectivity is part of the foundational infrastructure of AI. The Local Networks (LocNet) initiative, a collective effort anchored by the Association for Progressive Communications (APC) and Rhizomatica, serves as a powerful embodiment of the Ubuntu and Harambee spirit. LocNET is laying the ground for the basic infrastructure of AI to work in the global south by working with communities to build sovereign alternatives to connectivity that are community-centred.2 These initiatives are often called community networks but now carry the collective name of community-centred connectivity initiatives which thrive and embody the African ethos of ubuntu, ujamaa and harambee.
One of the innovative outcomes supported by this initiative was a microgrant for the development of a guide on “Offline AI” tools. In rural and remote areas where internet connectivity is sporadic or non-existent, these tools allow AI to run natively on local devices such as smartphones, desktops, or small PCs, without ever needing a cloud connection. Wakoma was supported to develop this guide that has introduced more ways for underserved communities to engage with AI including utilization of small language models. This project is not just a technical repository; it is a tool for digital liberation. It aligns perfectly with the goal of sovereign AI, allowing communities to build, optimize, and deploy AI models that are tailored to their unique local contexts, regardless of their proximity to a fibre-optic cable.3
In Kenya, this ethos was put to the test in the flood-prone Mavoko Sub-County of Machakos County (including Athi River, Kanani, and Jam City). Supported by LocNet, Digital Rurals strengthened community-centred connectivity by deploying an AI-enabled flood reporting platform integrated with IoT weather monitoring.4 This project embodies the African core principles of Harambee because it required the collective action of residents to build the mesh networks that carry these alerts. It also embodied Ubuntu by prioritizing the most vulnerable, ensuring that technology serves the collective security of the neighbourhood. Lastly it carried the principle of Ujima on collective responsibility by working with residents who became active participants in validating flood risks, turning local knowledge into actionable intelligence.
Case study in sovereignty: The Masakhane Network
To understand how these values are translated into practice, one need only look at the Masakhane Research Network. The name itself, “Masakhane”, is an isiZulu word meaning “we build together.” It is a pan-African, grassroots research network that has fundamentally altered the discourse around AI, shifting it from a top-down technological imposition to a bottom-up act of linguistic liberation.5
For decades, the digital landscape was a monolith of high-resource languages predominantly English, French, and Mandarin. When global tech giants developed AI, they utilized vast public and open internet data, which were overwhelmingly Western-centric. Consequently, African languages were largely ignored, given the status of “low-resource” languages. This was not merely an oversight; it was a form of algorithmic colonialism that effectively silenced millions of voices, ensuring that digital assistants, search engines, and automated educational tools could not interact with the nuances of local culture, idiom, or context.
Masakhane challenged this by mobilizing a decentralized collective of researchers, linguists, developers, and community members across the continent. Instead of waiting for Silicon Valley to include them, they began the process of curating their own datasets. Their methodology is an example of the Harambee spirit, pulling together resources, expertise, and time to achieve what no single entity could. They focus on Natural Language Processing (NLP) for languages like Luganda, Yoruba, Swahili, Wolof, and many others. Their work is a form of digital Kujichagulia (self-determination).
By building their own translation models and speech-recognition systems, Masakhane is ensuring that the digital future is not a mirror of Western values, but a reflection of African realities. They argue that if a language is not represented in the machine, it risks becoming obsolete in the digital economy. Through their research, they are proving that AI can be a tool for cultural preservation, ensuring that indigenous wisdom, oral histories, and local perspectives are not just stored, but understood by the technologies shaping our world.
Policy as philosophy: Institutionalizing African ethos
The transition from grassroots innovation to continental impact requires a robust policy framework. Africa is currently pioneering a governance model that embeds our traditional values into the digital regulatory landscape. The African Union’s Continental Artificial Intelligence Strategy (2024) serves as the guiding manifesto for this transition. Unlike global counterparts the AU strategy emphasizes developmental AI, technology designed to solve specific continental challenges in health, education, and food security..6 It explicitly advocates for the harmonization of data regulations across borders, a direct application of the Umoja (Unity) principle. By viewing data as a collective resource, the AU is working to prevent the digital balkanization of the continent. At the national level, countries like Tanzania, Rwanda, and South Africa are deploying localized AI strategies that echo this communal ethos:
Regulatory Sandboxes for Co-Innovation, echoing the Harambee spirit, nations like Mozambique are championing regulatory sandboxes. These allow startups, academia, and the public sector to test AI solutions in controlled environments, ensuring that innovation is monitored collaboratively rather than being stifled by bureaucracy. This mirrors the pulling together of expertise to mitigate risks while fostering growth.
Ujamaa-inspired Data Governance, Rwanda’s National AI Policy (2023) emphasizes data ecosystems that support public good. By focusing on frameworks like the AI4Health data-sharing initiative, Rwanda is moving toward an Ujamaa model where data is not hoarded for private profit but managed as a shared resource to improve public service delivery and agricultural productivity.
Kujichagulia-driven Data Sovereignty, emerging frameworks in South Africa and Tanzania are centring on human accountability. By mandating that high-risk AI deployments must maintain human oversight and utilize local data standards, these policies practice Kujichagulia. They ensure the intelligence governing public institutions is culturally congruent and ethically aligned with local societal norms, rejecting the one-size-fits-all standards imported from abroad.
The Ubuntu architecture: Coding for collective well-being
In the Silicon Valley model, AI is often designed to optimize for an individual’s needs, my personal shopping recommendations, my individual productivity, or my specific health metrics. Ubuntu-centred AI, however, views the individual as part of a web. If an algorithm is built to predict crop yields, an Ubuntu-informed model does not prioritize the high-tech farmer who can afford the most advanced sensors. Instead, it asks “How does this information benefit the local farming cooperative?”. It prioritizes tools that strengthen the village’s collective food security. It asks not how to make one person more efficient, but how to make the entire community more resilient against climate shocks.
The current global AI economy is built on a model of extraction, treating human data as a commodity to be harvested, sold, and used to train models that often serve the interests of distant corporations. Ubuntu stands in opposition to this. It views data as a communal asset. Under an Ubuntu framework, the data collected from a community such as health trends in a rural clinic or weather data from a local landscape, remains tethered to the community that provided it. It promotes data sovereignty, where the community has a say in how their digital collective is used, ensuring that AI serves their needs rather than being used to exploit them. Ubuntu emphasizes that harm to one member of the community is harm to all. In the context of AI, this provides a powerful framework for addressing algorithmic bias.
When a Western-trained AI system exhibits bias for example, misidentifying faces from certain backgrounds or making unfair credit assessments based on neighbourhood data, it is often treated as a technical bug. In an Ubuntu framework, this is a social failure. It is a failure to acknowledge the interconnectedness of society. Ubuntu-centred AI governance mandates that developers be held accountable to the real-world human impacts of their code. It forces a human-in-the-loop design where community elders, local experts, and those marginalized by the technology have the power to audit, critique, and veto systems. Finally, Ubuntu reframes inclusion not as a checkbox for social responsibility, but as a moral imperative. If we are all one, then a digital system that excludes the person with a disability, the rural farmer, or the speaker of a minority language is fundamentally broken.
This is why initiatives like Masakhane or the LocNet community networks are so powerful. They are not trying to add marginalized people to a pre-existing digital structure; they are building the structure from the perspective of those communities. Ubuntu teaches us that the quality of our future depends on the quality of our connections to the most vulnerable among us. If our AI systems do not work for them, they do not work at all.
In short, Ubuntu invites us to treat Artificial Intelligence not as a cold, objective tool, but as a member of our community. If we build our algorithms with the values of empathy, reciprocity, and shared destiny, we turn AI from a potential source of alienation into a powerful engine for collective flourishing.
The path forward
As we look toward the future of AI in Africa and the global south, the integration of AI must be seen as a profound moral choice. The journey we have traced from the linguistic decolonization led by Masakhane to the community-owned infrastructure by LocNet and the localized intelligence of the Offline AI projects provides a clear blueprint for Africa’s technological evolution.
To move forward, we must synthesize these lessons into a unified strategy:
Prioritize Sovereignty through Infrastructure: We must continue to treat power and the internet as foundational human rights. Scaling community-owned networks is not just a technical solution, it is a political act of Kujichagulia. By owning the pipes of the digital age, we ensure that the intelligence flowing through them remains accountable to local communities.
Institutionalize Ubuntu in Design: We must move beyond the “Move fast and break things” mantra. The future of African AI lies in “Move together and heal things”. We must mandate that every major AI rollout be it in healthcare, agriculture, or finance undergoes a social impact audit that measures success not by efficiency alone, but by how it bolsters the Ubuntu of the community it serves.
Invest in Frugal and Offline Intelligence: We must accept the reality of our current infrastructure and turn it into a strength. The development of offline-first, low-resource AI models should be a continental priority. This frugal AI creates a more robust, independent ecosystem that thrives regardless of the stability of international cloud connections.
Foster Regional Data Solidarity: Building on the Harambee spirit, we must establish cross-border data trusts. By sharing anonymized, ethically curated data across the continent, we can create AI models that understand the African context better than any external system ever could.
The movement toward sovereign AI is a testament to the resilience of the African spirit. By weaving the ancient values of Ubuntu, Ujamaa, Harambee, and Kwanza into AI, the continent is not just catching up; we are offering the world a new way to live with the machines we create. The servers are starting to hum, the community networks are growing, and the story of Africa’s future is, at long last, being written by Africans themselves.
Notes
1. Karenga, M. (1966). The seven principles of Kwanzaa: A collection of essays on the African cultural position. US Organization.
2. Association for Progressive Communications & Rhizomatica. (2025). Meaningful community-centred connectivity: The LocNet initiative. https://www.apc.org/en/partner/rhizomatica
3. Wakoma. (2025). OfflineAI: Local/offline machine learning resources [GitHub repository]. https://github.com/Wakoma/OfflineAI
4. Digital Rurals. (2025). Community-centred connectivity and climate resilience in Mavoko Sub-County: A case study of IoT and AI-enabled flood reporting [Video]. YouTube. https://www.youtube.com/watch?v=k3bEtJbwbdg
5. Masakhane Research Network. (2026). Building inclusive AI for African languages: About us. https://www.masakhane.io/
6. African Union. (2024). Continental artificial intelligence strategy for Africa. https://au.int/fr/node/43871
Rebecca Ryakitimbo is a feminist technologist, researcher, and community leader working at the intersection of artificial intelligence (AI), language data, gender justice, and digital equity. With over a decade of experience spanning research, grassroots networks, and global fellowships (Google, Mozilla, ISOC, National Geographic society). She supports feminist tech spaces such as the African Women School of AI and curates the Gendering AI conference. As part of LocNet, she supports CCCIs by facilitating Communities of Practice, and researching community-centered connectivity and local services for equitable, locally led digital ecosystems.