One of the challenges that the group identified is the lack of visibility of existing tools that help AI ecosystem actors operationalize AI fairness in their institutions or processes. A list that could grow through an ongoing crowdsourcing effort. Thus the idea of a global library was born: a digital space that builds upon existing efforts and unites key reports, resources, tools, and methodologies on the subject across geographical and linguistic barriers. A space that makes it easy to keep track of the latest developments in this field and access actionable information in an agile way.  

We are profoundly grateful to the World Economic Forum team which provided direct support to every step of this effort and to the women-led action tank, C Minds, which works on the intersection of new technologies, society, and the environment, and joined as a partner to support the development of this Library. Moreover, we are thankful for the inspiration we had from other efforts such as UNICEF's mapping of resources on AI for Children, and appreciate deeply the people that participated in the crowdsourcing effort to map key resources analyzed during the curation phase. 

​

The first iteration of the AI Fairness Global Library is a list curated by members of the Council and is by no means exhaustive. This initial list serves as a kick-off effort and can be complemented with more entries through collective action. The library has the secondary aim of serving as input to other important efforts advancing the field of AI governance, such as the Global AI Action Alliance (GAIA), the OECD's framework to compare implementation tools, UNESCO's AI ethics global instrument network, IDB's fAIr LAC observatory, and the Japanese Repository, among many others that could benefit from this mapping. 

​

Our understanding of the resources collected in this Library is further informed by interviews with key stakeholders in the AI governance ecosystem. Major challenges identified during interviews included the time and material investments necessary to achieve meaningful AI fairness outcomes; and the difficulty of meaningfully bringing disenfranchised communities into mainstream conversations and companies, especially ensuring they achieve leadership positions. We hope this Library is a step toward addressing both challenges and most importantly creating an inclusive and open community of stakeholders working toward the responsible development and use of AI systems.

​