ITU's GENIE.AI framework update highlights public-sector AI tool for low-resource governments

An article reports on a 22 October update from the International Telecommunication Union (ITU) about GENIE.AI, a framework meant to help governments build and run generative AI services while keeping more control over technology, data and infrastructure. ITU describes the platform as modular, open source and model-agnostic, aimed mainly at public bodies with limited resources, technical capacity and expertise.

The report places the update in a longer timeline. Work began in 2024, a high-level reference architecture was adopted in January 2025, and the platform, ITU says, formally launched in July 2026, so the October item is not a new launch. Instead it shows the effort moving from technical development toward real deployment in the public sector.

According to the article, institutions can use the framework to create chatbots, digital assistants and content-generation tools, while avoiding dependence on one commercial vendor. It carries no licensing fees, works with GovStack digital public infrastructure, and can run in the cloud or on premises, including where large-scale infrastructure is scarce. ITU's framework page cites over 100 active early adopters alongside 10 engaged countries and 30 real-world use cases spanning agriculture, health, education, climate and digital public goods.

Country examples include Kenya, where ITU has helped state agencies and homegrown developers acquire open-source AI know-how, and El Salvador, where GENIE.AI underlies AgroGenio, a farm advisory prototype whose knowledge base draws on 150-plus guidance documents from national and international sources. Public-sector uses in Bangladesh, The Gambia and Lesotho have received backing through the ITU-IEEE GenAI for Good Challenge. The programme ties to ITU's initiative that enables open source ecosystems and to broader efforts around digital public goods and infrastructure that interoperates.