Solution

Shaping AI for the public interest

The AI for Development initiative envisions a world where ethical and inclusive AI applications contribute to the public good. The initiative will fund AI innovations, research to inform use and policy, and capacity building activities. The driving research entry points will look at: infrastructure and skills gaps and opportunities, future of work, policies and regulations for an inclusive and rights-based AI, ethical AI applications, and impact of AI on human and democratic rights. (See white paper on AI and human development for a detailed research agenda, or annex for a summary.)

The targeted changes or outcomes and the indicative key activities of the program can be organized into the three levels that constitute AI ecosystems: policy and regulation, applications, infrastructure and skills.

Improve policy and regulation to promote inclusive and ethical AI

Outcomes:

  • Greater AI expertise built within developing country governments, including regulatory entities, new advisory bodies, and funding agencies
  • Evidence-informed AI regulatory models adopted by governments to deal with the emergence of new AI-driven activities such as predictive policing and autonomous vehicles
  • AI ethics guidelines developed and adopted by countries

Indicative key activities:

  • Training to government officials, including peer networking and open educational resources adapted to local context.
  • Global survey of government’s AI readiness and implementation, benchmarking leaders in the ethical use of AI
  • Core support for multi-disciplinary research to policy centres to engage in policy work on intersection of AI related innovation, protecting human rights, and security

 Spur AI applications in the public interest

Outcomes:

  • Applications that address pressing social, economic and environmental issues
  • Better understanding of the intended and unintended social implications of AI technologies in the Global South
  • Better practices for participatory AI design to support the development of inclusive AI applications
  • Deeper understanding of the process of scaling cost-effective and inclusive AI applications
  • Peer-reviewed publications exploring new AI for development techniques and documenting the use of impact of AI in society

Indicative key activities:

  • Annual competition to seed inclusive AI applications in the public interest
  • Innovation accelerators to support the scaling of evidence-proven AI applications (selected annually)

Foster AI infrastructure and skills in the Global South

Outcomes:

  • New or strengthened multidisciplinary AI labs within universities & colleges to enable locally relevant and locally designed AI applications
  • Strengthened, accessible AI data infrastructure in the global south
  • Reduction in AI skills gap at all levels of design, development, and deployment, with a focus on women and other marginalized populations
  • Strengthened capacity and reputation of multidisciplinary AI research to policy work, through research institutions engaged in activities at national, regional, and global levels

Indicative key activities:

  • Supporting research chairs and interdisciplinary labs in machine learning for the public interest
  • Training of southern research councils in funding ethical and inclusive AI research
  • Development of research informed open, online training curriculum and courseware on ethical and inclusive AI, and AI governance from a Southern perspective
  • Contributing to the creation of open access, global south relevant data sets
  • Fellowships and/or training programs targeted to specific skill development

 

There will also be a set of activities that contribute across all three areas, including:

  • Bi-annual baseline global survey of AI capacities, use and implementation across the global south
  • Southern-led regional and thematic research networks focused on issues such as AI and education, future of work, data governance, biases and inclusion.
  • Series of regional and thematic events and an annual “AI in the Global South Conference”