Artificial intelligence for development cooperation—FAIR Forward

Remove entry barriers to AI: FAIR Forward facilitates the provision of open, non-discriminatory and inclusive training data, models and open-source AI applications. This includes open language data to enable voice-based access to services and information, and earth observation data to empower marginalised groups with access to climate-smart agricultural advice. Through these efforts, we aim to ensure that AI benefits everyone.

As part of this effort, FAIR Forward has worked on several key projects, including:

• Using AI to combat climate change and tackle deforestation
  
  In collaboration with HCS Foundation (HCSA) and local mapping organization JKPP, we gathered important information about trees throughout Indonesia’s forests, that can improve the accuracy of forest classification using artificial intelligence. The project is now being scaled to India and Senegal.
  With Lacuna Fund, we have been supporting the creation of datasets for machine-learning applications for under-resourced languages and agricultural use cases since 2021. Through the Climate and Forests call, 14 projects across Asia, Africa and Latina America will create vital datasets for climate resilience and forest preservation.

• Language technology for underrepresented languages
  
  FAIR Forward works to address the shortage of AI training data in emerging and developing countries. For this purpose, we work closely with many partners, including the Rockefeller Foundation, Google.org, Gates Foundation, Canada’s International Development Research Centre and Masakhane.
  With Mozilla Foundation, FAIR Forward has been working towards building open voice datasets and support voice-enabled use cases for low-resource languages in Africa and Asia.
  In India, together with the Indian Institute of Science (IISc), FAIR Forward is collecting language data in ten Indian languages and making them freely and openly available to Indian partners. Overall, our efforts to open language technology in under resourced languages spans African, Indonesian and Indian languages, widely spoken and benefiting over 600 million people worldwide.

• AI for smart and resilient agriculture
  
  Using hyperlocal geo-referenced data and high-resolution satellite imagery, FAIR Forward supports climate-smart agricultural use cases. In Kenya, with the Local Development Research Institute (LDRI) we developed an AI Early Warning System that monitors farming activities and supports precision agriculture practices in Kiambu and Embu counties of Kenya.
  In the Indian state of Telangana, we have partnered with the Department of Agriculture & ITE&C Department to support the development of an AI-based crop type map for the state.

• Datasets
  
  Since 2019 FAIR Forward has worked with its partners in Africa and Asia to create open AI datasets to address local challenges. For a full overview explore our growing GitHub page and learn more about the 20+ datasets and AI models available as open-source assets for local AI communities.

FAIR Forward supports digital learning and training for the development and use of AI and fosters cooperation with German and European research institutions and businesses. These efforts have brought AI expertise to over 120,000 professionals across our partner countries.

As part of this activity/action-area, FAIR Forward has worked on several key projects, including:

• Global E-learning Courses on Open AI
  
  Joining forces with Makerere University Uganda, FAIR Forward has developed a 5-part e-learning course on creating open and unbiased datasets for artificial intelligence across different domains. The courses are available for free to anyone on the atingi e-learning platform and can be customised for local adaptation with the help of a dedicated replication kit.

• AI Bootcamp for Women
  
  FAIR Forward created the AI and Data Science Bootcamp for Women and Minorities. The first edition of the AI and Data Science Bootcamp for Women and Minorities was held in South Africa, through a partnership with the BMZ initiative Data Economy and Intel. An adjusted version, the AI Fellowship for Women was also successfully carried out in Ghana, in partnership with Blossom Academy, and in Rwanda, with the Ministry of Information, Communication and Technology.

• AI Skills for Professionals
  
  The Department of Communications and Digital Technologies (DCDT), supported by NEMISA, in collaboration with FAIR Forward, developed an AI training programme to equip young South Africans with the skills they need to develop local AI solutions. This knowledge enables them to solve local problems and contribute to economic growth in the South African AI industry.

FAIR Forward advocates for value-based AI that is rooted in human rights, international norms such as accountability, transparency of decision-making and privacy, and draws on European regulation such as the EU AI Act and the General Data Protection Regulation (GDPR).

Some key projects in these areas include:

• AI regulation—developing policy blueprints and national AI strategies
  
  Smart Africa championed the development of the Pan-African AI policy blueprint “AI for Africa” with the support of FAIR Forward. Our project has supported the development of national AI strategies in our partner countries, including Rwanda, Kenya, Ghana and South Africa.

• Asia-Africa Policy Makers Network
  
  Launched in 2022, the Africa-Asia AI Policymaker Network offers a valuable platform, enabling collective and cross-country exchange of government officials from Africa and Asia on policy frameworks and strategies to develop responsible and locally relevant AI applications.

• Responsible AI Assessments (AI Governance in Action)
  
  FAIR Forward, along with Eticas has developed a method to conduct a holistic AI risk and ethics assessment. The method bridges the gap between abstract, high-level AI ethics principles and practical implementation by providing a structured approach to operationalize these principles, enabling AI stakeholders to identify and mitigate potential risks effectively.