As the promise of artificial universal intelligence (AGI) increasingly captures global imagination, it is crucial that we ensure that AI is moving forward, not only to give everyone a privileged community, not only to those already have relatively wealthy resources, but especially to underserved people facing ongoing educational and economic disparities. From our experience working together at the ICOG Laboratory in Ethiopia, a company co-founded by Ben Goertzel and GetNet Aseffa, the first in Ethiopia in 2013 and remains the most important AI company to date, we have witnessed first-hand the challenges of applying AI technology to developing AI technologies in a developing world.
AI’s potential as an educational equalizer is far-reaching. But for many communities, especially outside major urban centers, or struggling to cope with huge socio-economic barriers, even basic quality education is elusive. Above many other challenges in life in developing countries, these underserved populations often encounter two core challenges specific to the education field: language barriers and culturally insignificant educational content. These can be overcome, but we find that doing so can require a lot of art and sufficient resources, especially the need to understand the technology itself and the specific local difficulties faced in the case of developing the world.
Overcome language barriers
UNESCO estimates that 40% of students worldwide are unable to receive education in languages they fully understand. It doesn’t take much imagination to see how this fundamental disconnect can severely hinder learning. However, AI-powered translation and language tools provide powerful solutions. This is one of the ways advanced technology can provide huge benefits to underserved populations relatively inexpensively. However, the developed world tech companies that drive most modern AI development have little motivation to perfect language technology, mainly spoke by individuals with minimal purchasing power, no credit card, and little chance or tendency to click on ads.
The collaboration we have developed between the ICOG Lab and Curious Learning embodies the potential here. Using the generated AI, we have carefully crafted local native language reading applications, currently serving more than 85,000 active users. Such initiatives show how AI can help overcome language barriers, even low-resource languages served with standard large language models.
We recognize data scarcity as a bottleneck, and we also launched Leyu, a decentralized data library platform that explicitly collects language resources from disconnected communities. Local AI developers can then use the collected data, such as sentences in a pair of languages and popular languages with insufficient resources and better resources, to train the AI model that transforms the local language into the world’s most Internet-based world languages. By proactively addressing this language gap, we ensure that communities benefit immediately when connected, rather than lag further.
Ensure relevance through contextual learning
In addition to language, effective education requires relevance. Imported educational content often does not resonate with learners, whose daily experiences vary greatly from the programs described in standardized courses. AI enables customization of educational materials and contextualizes local realities courses. Imagine a mathematical problem derived from science education using local agricultural practices or community market transactions. This culturally consistent content is not just about education, it inspires practical applications, fosters participation and self-reliance.
Our Digitruck Project is an off-grid mobile education center deployed by ICOG Labs and partly sponsored by our globally decentralized AI Project SingularityNet, which vividly demonstrates this. We have equipped a semi-trailer truck as a portable classroom with computers and electronics and brought it to a local community composed of local expert teachers. Young learners in rural Ethiopia have concealed the ability of AI to implement other technologies by encountering coding and artificial intelligence concepts through hands-on experience using tablets and manufacturer toolkits and applications in relevant environments such as improving agricultural practices.
Responding to the diversity challenges posed by developing world ecosystems can require considerable patience. For example, during the 2015-2019 period, our Robosapiens initiative introduced AI to students at the University of Ethiopia through humanoid robots that program football playback, a culturally resonant and engaging approach. The robot football match between Ethiopians, Kenya and Nigerian universities has a fierce energy for participating students, which is frustrating when we have to pause the program, due to the complexity of the high import tariffs associated with electronic promises that cannot be obtained even in the government itself (part of itself).
AI is a trustworthy ally, not a threat
Contrary to the fear of an increasing number of digitally saturated societies such as Terminator Risk or Work Displacement caused by AI, communities with limited internet access often view AI in a different way: as trusted information allies. For example, Nigerian farmers actively work with AI-backed call centers to gain practical agricultural advice and market insights. Here, AI technology complements and enhances rather than threatening livelihoods, thereby enhancing trust through tangible benefits.
Support collective learning and social structure
AI integration into education must respect the existing social structure. Many underserved communities prioritize collective over individualistic approaches, making group learning crucial. Helpful AI should foster collaboration, enhance community guidance, and integrate seamlessly with existing collective decision-making processes. Tools designed from a decentralized and participatory perspective naturally align with this community-driven educational model, thereby strengthening rather than undermining social cohesion.
As a concrete example of how it works, one can envision extending the Digitruck initiative into a more lasting program where Digitruck alumni guides integrating AI into every aspect of rural life in Ethiopia. We hope that the AI-powered education platform blends with community-led workshops. Imagine community elders and teachers working together to use AI-generated learning materials during group meetings to promote discussion on practical topics such as sustainable agricultural technologies, local health care practices and financial knowledge. These AI tools will not simply provide content; they will actively encourage group dialogue and collective problem solving, strengthen community bonds, and ensure that education remains deeply embedded in local traditions and collective decision-making frameworks. This kind of program will be simple enough to deploy now; all that is missing is to plan for this “only” funds.
Navigation risk and ethical implementation
The commitment to accelerating the development of the world’s positive self-transformation is obvious and exciting, but nonetheless, we must address these risks. The ease and immediate risk of AI reduces students’ basic skills or motivation. The introduction of AI responsibly requires strengthening rather than replacing human educators and traditional learning foundations. AI must be positioned as a supportive infrastructure – promoting personalized learning and stimulating intellectual curiosity, rather than answering generators, to undermine critical thinking and motivation.
As we progress in these directions, it is crucial to pay close attention to human alignment for very practical reasons: if it does not meet the needs and values of the local population, AI will not provide the services needed to those who need it most. However, we strongly believe that consistency should emerge from rich and meaningful collaboration, rather than rigidity and ham hand rails. Rather than limiting AI in narrow, predefined values, is mapped by boundaries controlled by a particular culture or elite, but rather stems from experiences of real participation, whereas AI has a deep connection with human learners. This is how people shape humans and artificial intelligence systems, thereby promoting mutual growth.
Decentralization and Democracy of Global Education
We have already suggested that a handful of large companies from two major countries currently dominate the global AI technology field. This domination is the core reason AI language technology currently ignores most African languages, and is often more useful than rural poor people in Africa, Central Asia, or elsewhere.
While we respect the amazing work these large tech companies have done, we firmly believe that decentralized, democratic guidance on the development of AI development is a key advantage of global educational equity. That’s why we put so much energy into developing platforms like SingularityNet, which can decentralize AI architectures and empower governance based on broad participation and democratization. Such a framework makes AI development more likely to reflect multiple global needs rather than narrow corporate or government interests.
We understand that the path to equitable AI-enhanced education is not a straightforward one – requiring intention, cultural sensitivity, moral vision and participatory governance. But the potential rewards – removing educational barriers, enhancing cultural significance, and empowering communities around the world – make this journey not only worth it, but urgent.
Through careful management, we can use ever-increasing artificial intelligence to achieve educational equality and generally improve humanity. These sound like abstract Aspergillus albicans, but when a child writes their first line of AI code in a Digitruck visiting their village, their specific meaning is clear.
