Nội dung text GIRAI Methodology and Conceptual Framework -2.pdf
Methodology and Conceptual Framework of the Global Index on Responsible AI. 1st Edition
Building the Global Index on Responsible AI The structure and content of the Global Index on Responsible AI (GIRAI) emanates from a series of consultations with a wide range of stakeholders from around the world, specifically those from Africa, Asia, the Caribbean and South and Central America. During these consultations, participants were invited to engage with the scope and structure of the GIRAI, which not only included the conceptual foundation and rationale for the Index, but also areas of measurement, including the three pillars, three dimensions and thematic areas. Participants were invited to comment on the suggested areas of measurement and the interpretation thereof, as well as the approaches to data collection, analysis, and reporting. Specific attention was paid to relevance of each thematic area across countries, as well as the local prioritisation of different human rights groups and the potential availability of data in the country. In September 2023, a pilot of the GIRAI was completed in 10 countries, including Costa Rica, Kenya, Jamaica, Canada, Serbia, India, Sri Lanka, Palestine, Burkina Faso, and Georgia. Preliminary findings from the pilot study validated the Global Index’s methodology and demonstrated its proof of concept. The countries included in the 1st Edition of the Global Index on Responsible AI were chosen based on where suitable country-researchers were recruited and the existing research network of institutional partners. Governments were not involved in determining whether their country was included in the Index. Data was collected for Chinese Taipei (commonly referred to as Taiwan) and the West Bank, which is under the administration of the Palestinian Authority. No data collection took place in Gaza. We recognize that these are, at the time of writing, disputed jurisdictions. 3
Index structure The Global Index on Responsible AI is composed of: ● 3 dimensions that align with human rights-based standards and democratic principles, ● 3 pillars accounting for the AI ecosystem, ● 19 thematic areas covering core components of the concept of Responsible AI and, ● 57 indicators measuring the performance of each thematic area within each pillar. Dimensions The Global Index on Responsible AI is composed of three Dimensions rooted in human rights and democratic principles, these are: Responsible AI Governance This dimension measures the degree to which country-level governance regimes uphold effective and rights-preserving practices in responsible AI. 4