Shaping a safe, equitable and inclusive future with AI

People use artificial intelligence (AI) every day to help make decisions, solve problems and automate tasks. Although AI applications have played a growing role in our daily lives for years, they are now in the spotlight after GPT-3, a language model that is able to perform an impressive range of natural language tasks, was launched in late 2022. Now more than ever, people are asking how societies can ensure AI will improve life for everyone, regardless of where they live. Research can help find these answers. 

The AI story at IDRC 

IDRC is one of the largest funders of AI research in the context of international development. The research we support is centred around the concept of responsible AI: the practice of designing, developing and deploying AI systems that are ethical, inclusive and sustainable. IDRC-supported research is also primarily carried out by scientists in the Global South because IDRC knows that people who are closest to a challenge are best placed to develop solutions.

In 2017, Canada became a leader in AI as the first country to announce a national strategy, the Pan-Canadian Artificial Intelligence Strategy. IDRC followed suit in 2018, producing the Centre’s white paper on Artificial intelligence and human development. This proposal for a concrete research agenda for responsible AI in the Global South forms the basis of IDRC’s current programming approach. This article provides an overview of much of that programming. 

AI4D Africa: Strengthening the AI ecosystem in Africa 

The Artificial Intelligence for Development Africa (AI4D Africa) initiative is a five-year, CAD20 million program designed to promote the use of responsible AI to address problems and improve lives in Africa. Co-funded by the Swedish International Development Cooperation Agency (Sida), the AI4D Africa program collaborates with and funds research by the AI practitioner, science and policy communities across 19 countries in sub-Saharan Africa. 

AI4D Africa focuses on three pillars:  

• Developing, testing and scaling responsible AI innovations that address Africa’s development challenges through five cross-regional innovation research networks (focused respectively on climate action; agriculture and food systems; gender and inclusion; sexual, reproductive and maternal health; and education and supporting innovative small-grant programs). 
• Supporting policy research to inform and facilitate the development of public policies and regulations that promote the inclusive benefits of AI by funding two AI4D research-to-policy think-and-do tanks in anglophone Africa, a policy network in francophone Africa, and an African Observatory on Responsible AI. 
• Building capacity in academic institutions and among researchers to foster African talent and skills around AI by developing three multidisciplinary AI labs in public universities and the African AI4D Scholarships Program. 

In only a short amount of time, IDRC and its partners have strengthened the burgeoning AI ecosystem in Africa, fortifying the already-rich landscape with increased capacity to develop homegrown, responsible AI for the benefit of generations to come. Some early outputs include a paper on children’s privacy concerns, a study on the impact of AI on women's job losses in Africa’s banking industry and a set of learning briefs for researchers focused on gender-responsive AI.  

Feminist AI: Advancing gender equality and inclusion in AI research 

Established in 2021, this CAD2 million project has two main goals: to improve how AI can address social and economic barriers faced by women and marginalized communities and to correct historic exclusion and bias that currently exists in AI systems.

A global research network supports research hubs and projects across Asia, the Middle East and North Africa, and Latin America, and coordinates with AI4D Africa. Advancing feminist thinking in AI spaces and platforms has the potential to be transformative and is therefore a critical area of investment for IDRC.

Responsible AI for health 

At the outset of the COVID-19 pandemic in 2020, IDRC and Sida collaborated to fund the Global South AI4COVID Response Program. The program’s eight projects aimed to strengthen health systems and improve pandemic response in countries across Asia, Latin America and sub-Saharan Africa. AI4COVID research led to evidence-informed COVID responses in several countries, including the development of official dashboards viewed by thousands of people every day in sub-Saharan Africa; recommendations on the return to school in Colombia; and contributions to the WHO guidelines on the ethics and governance of AI for health.   

IDRC’s current focus in the AI and health sector is the Artificial Intelligence for Global Health initiative. This CAD15.5-million investment, launched in 2022, is supporting research in three key areas: epidemic and pandemic prevention, preparedness, and response; AI innovation research networks on sexual and reproductive health in Asia, the Middle East and North Africa and Latin America; and the role of the private sector in AI and health innovation in Africa, Latin America and Asia. 

Other work we support in Responsible AI: 

• The Global Index on Responsible AI: The index aims to measure and bolster the capacity of governments, civil society organizations and other stakeholders to uphold responsible AI principles by benchmarking their progress over time. The intention is for the index to influence policy and practice at a national scale by providing policymakers with independently verified data to help them plan how to better foster accountability, assess the impact of AI use, and encourage regional and international cooperation and regulation on AI issues. Work is in the early stages but the project is already garnering attention. It was selected for the Scale-up Program at the Paris Peace Forum 2022 to receive year-long customized, non-financial support in 2023.    
• Lacuna Fund: IDRC, alongside Google.org and the Rockefeller Foundation, founded the Lacuna Fund to address the data gap in marginalized contexts. This is the world’s first collaborative effort to provide data scientists, researchers and social entrepreneurs in low- and middle-income contexts globally with the resources they need to produce labeled data sets that address urgent problems in their communities. Launched in July 2020, the Lacuna Fund has grown from USD4 million to USD15 million, raising funds to create machine learning datasets that would help address development challenges in areas such as agriculture, languages, health and climate. 

Working for the future 

The AI revolution is just beginning. Our research partners are already exploring how to harness AI to address poverty and exclusion, improve food systems, confront the challenges of climate change, and make education more inclusive. Together we have learned that it is important to temper enthusiasm with a critical understanding of the potential risks of AI systems.

The emergence of the new large language models such as ChatGPT make this work even more urgent. For AI to bring developmental benefits, it is critical that sufficient local capacity, infrastructure, and evidence are in place to inform government strategy and policy. That is why we are helping to build an ecosystem of AI research and innovation that can adapt to new and potentially even more disruptive AI applications in the future.