Analysis of the Global Index on Responsible AI (GIRAI) Methodology and Recommendations for Improving Uzbekistan’s Position
The Global Index on Responsible AI (GIRAI) was first published in mid-2024 and assesses national progress across 138 countries on the ethical and human-rights-respecting deployment of artificial intelligence (AI). The Index covers 19 thematic areas across 3 pillars: Government Frameworks, Government Actions, and Non-State Actors’ Initiatives.
The Index documents define Responsible AI as follows:
Responsible AI refers to the design, development, deployment, and governance of artificial intelligence in a manner that respects and protects all human rights and adheres to ethical principles throughout the entire AI lifecycle and value chain. It requires all actors within national AI ecosystems to take responsibility for the human, social, and environmental impacts of their decisions.
Similar indices include the Government AI Readiness Index published by Oxford Insights, as well as newer global measures such as the newly launched OECD.AI Index and the IMF/World Bank AI Preparedness Index (AIPI), which covers 174 countries.
A comparative overview of currently active Global AI indices is provided below:
| Name | Developed by | Focus (Primary Orientation) |
| Global AI Index | Tortoise Media | AI capacity |
| AI Index and Global AI Vibrancy Tool | Stanford Institute for Human-Centered Artificial Intelligence | AI achievements and AI vibrancy |
| AI Preparedness Index (AIPI) | International Monetary Fund (IMF) | Countries' readiness to adopt AI |
| AI Readiness Index | Oxford Insights | Public sector readiness |
| Global Index on Responsible AI (GIRAI) | Global Center for AI Governance | Progress toward responsible AI. Ethics and human rights. |
| AI and Democratic Values Index | Center on AI and Digital Policy | Alignment of AI policy with democratic values |
Table 1. Comparative overview of Global AI indices
Scoring Methodology
In its inaugural 2024 edition, the Index assessed 138 countries using a rigorous methodology combining primary and secondary data sources, human rights-based criteria, and a multi-stakeholder evaluation framework.
The Index uses a hierarchical analytical structure organised around three dimensions, nine sub-dimensions, and nineteen thematic areas. These thematic areas cover the social, technical, and political aspects of responsible AI governance, enabling the Index to capture the full scope of national AI ecosystems rather than limiting itself to normative documents alone. Results for each thematic area are assessed across three pillars:
Government Frameworks assesses the formal legal and policy architecture, including national AI strategies, dedicated legislation, and regulatory structures.
Government Actions assesses concrete government initiatives aimed at advancing responsible AI in practice, acknowledging the experimental and dynamic nature of AI policymaking beyond formal commitments.
Non-State Actors’ Initiatives captures the contributions of civil society, academia, private sector entities, and other non-state stakeholders to the responsible AI ecosystem.
1. Primary Data Collection
The primary data underpinning the Index were collected through a large-scale expert survey conducted by in-country researchers between November 2023 and February 2024. Experts across all 138 countries completed a comprehensive questionnaire consisting of 1,862 questions (98 questions per thematic area).
Country selection was determined not by government bodies but by the availability of qualified local researchers and inter-institutional collaboration, which safeguards the Index’s independence.
Following data collection, a global quality review team systematically examined all submitted data for consistency, accuracy, and completeness. This quality assurance stage is an essential part of the methodology, reducing researcher subjectivity and ensuring that cross-country comparisons rest on reliable evidence.
2. Score Calculation, Secondary Data, and Aggregation
Index scores are derived primarily from expert survey data. Scores for each thematic area are calculated separately for each of the three pillars, normalised to a 0-100 scale, and then averaged to produce pillar-level scores.
These primary data scores are subsequently adjusted using a coefficient drawn from secondary data sources for each country. Once the adjustment is applied, each pillar is assigned a specific weighting coefficient, and the results are aggregated into a single overall score for each country.
The Index’s open data policy means that country scores and the underlying evidence are available for independent verification and future research.
Uzbekistan’s Current Standing in GIRAI
(The GIRAI report was last published in 2024)
1. Overall Score: 11.27
Uzbekistan scored 11.27 out of 100, the highest result in the Central Asian region (neighbouring Kazakhstan, for instance, ranks 74th on this list). In the broader global GIRAI context, two thirds of all countries scored below 25 points. The score reflects the gap between Uzbekistan’s rapid pace of high-level policymaking and the quality of implementation, inclusiveness, and rights protection that the Index directly measures.
2. Pillar Scores
Government Frameworks: 14.21
Uzbekistan’s highest score across all pillars is linked to key documents such as Presidential Resolution No. PQ-358, which sets targets including growing AI-based services revenue to $1.5 billion by 2030 and breaking into the top 50 of the Government AI Readiness Index.
Government Actions: 12.00
The Index recognises Uzbekistan’s active development of AI across healthcare, education, agriculture, fintech, energy, and taxation. However, the relatively low score reflects limited centralised accountability mechanisms and restricted independent oversight.
Non-State Sector: 3.95
This low score reflects the broader systemic state of the governance environment in Uzbekistan, indicating that the engagement of the private sector, civil society, and academic communities on responsible AI remains very weak.
3. Dimension Scores
Responsible AI Capacity: 24.47
The strongest dimension, shaped by government investment in workforce training, computing infrastructure, and AI research institutions. While these capabilities remain concentrated primarily in urban centres, the overall development trend in the field is positive.
Responsible AI Governance: 11.15
This score reflects the absence of a dedicated AI regulatory body, an algorithmic impact assessment system, or transparency obligations for government AI systems.
Human Rights and AI: 5.78
The weakest dimension, assessing the degree to which fundamental rights — including privacy, freedom of expression, non-discrimination, and access to legal remedies — are respected and protected in the development and deployment of AI. This score reflects a deep systemic tension between Uzbekistan’s ambitious AI objectives and its human rights record.
Conclusions and Recommendations
To further develop Uzbekistan’s AI ecosystem and strengthen its position in global rankings, particularly in the GIRAI, the following set of concrete measures is recommended based on identified needs:
1. Strengthen the Legal Framework
To address the need for dedicated legislation establishing accountability and safety standards for AI algorithms, it is advisable to adopt an AI Governance Law within the next 1 to 3 years. While this process requires significant resources, it will regulate the ethical and legal norms of the field.
2. Adopt International Ethical Standards
To ensure alignment with global principles, formally localising UNESCO and OECD AI principles within 12 months is a practical step. This initiative requires relatively few resources and establishes a strong foundation of trust in the field.
3. Build Scientific and Institutional Capacity
To enrich the research base and expand the pool of highly qualified professionals, funding AI research centres and introducing dedicated PhD grants over the next 1 to 3 years is recommended.
4. Expand Digital Infrastructure
To meet demand for high-performance computing resources and bring AI projects into practical application, it is important to establish Sandbox environments for government AI projects and modern cloud computing systems within 1 to 3 years.
5. Raise AI Literacy in Public Administration
To build effective AI skills among civil servants, introducing dedicated literacy courses within 12 months is recommended. This will in turn improve digital government efficiency and a culture of working with open data.
In summary, Uzbekistan’s strength lies in political will and strategic planning. However, human rights-related aspects and the limited role of the non-state sector continue to weigh down the country’s overall ranking. To reach its target indicators by 2030, the primary focus must be directed toward technological infrastructure, scientific capacity, and data governance.
References
- Global Index on Responsible AI 2024 Report: [Link: global-index.ai]
- GIRAI Methodology and Conceptual Framework: Global Index on Responsible AI. [Link: Google Drive]
- Government AI Readiness Index 2025 Report: [Link: oxfordinsights.com]
- International Monetary Fund (IMF). “AI Preparedness Index (AIPI) Dashboard”. [Link: imf.org]
- OECD AI Observatory. “OECD AI Observatory Index: A comprehensive view of national AI ecosystems”. [Link: oecd.org]
- UNESCO. “Readiness Assessment Methodology (RAM): Ethics of Artificial Intelligence”. [Link: unesco.org]
- Institute for Development of Freedom of Information (IDFI). “Responsible AI in Eastern Europe and Central Asia: Regional Analysis of GIRAI”. [Link: idfi.ge]
- Law of the Republic of Uzbekistan, No. LRU-850 dated 03.07.2023 “On Competition”. [Link: lex.uz/uz/docs/-6518381]
- Official resources of the Competition Promotion and Consumer Protection Committee of the Republic of Uzbekistan. [Link: raqobat.gov.uz]
