CIAIRN: Catalyzing inclusive AI research network
Programs and partnerships
Lead institution(s)
Summary
Artificial intelligence systems trained on datasets that reflect historical biases or lack diversity can perpetuate existing inequalities.Read more
Artificial intelligence systems trained on datasets that reflect historical biases or lack diversity can perpetuate existing inequalities. For example, facial recognition technologies have demonstrated higher error rates for individuals with darker skin tones due to under-representation in training data. Similarly, algorithms used in hiring processes may put certain groups at a disadvantage. A 2023 study published in the journal Nature found performance disparities in large language models across different languages and dialects, which could affect speakers of less-represented languages.
AI systems built on data that embed existing harmful social norms tend to perpetuate these biases and produce unfair outcomes, while excluding vulnerable communities. Mainstream AI development happens in a science, technology, engineering and mathematics (STEM) system that continues to be unequal, with far fewer women and people with disabilities represented.
Generally driven by market forces, AI systems seldom prioritize the requirements and experiences of communities and vulnerable populations. They offer little insight into how decisions are made, who is accountable, or who is being excluded. While the principles of responsible AI are generally understood, their implementation remains limited, particularly in weaker regulatory environments.
The Catalyzing Inclusive AI Research Network (CIAIRN) aims to catalyze inclusive, human rights-based and community-driven AI innovation to address the problem of how AI systems replicate harmful gendered social norms and, instead, develop solutions that ensure the benefits of AI are distributed equally.
This project aims to create AI solutions through community-centered innovation, ensuring socially just and responsive technologies. Experts will collaborate with communities to develop adaptable AI tools that meet their needs. The initiative will build community capacity to understand AI development and governance, and share knowledge among developers, policymakers and the public. CIAIRN is a global network, including institutions from across Latin America and the Caribbean, Asia, sub-Saharan Africa, the Middle East and North Africa, plus a global coordinating body to consolidate knowledge and support the growing network.
About the partnership
Artificial Intelligence for Development
Share this page