GIRAI 2024: notes from the Kyrgyzstan national researcher chair

On 14 June 2024, the first edition of the Global Index on Responsible AI (GIRAI) was published. Kyrgyzstan placed 79th out of 138 countries assessed. The index is publicly available at https://global-index.ai. I want to write this post from a specific seat — I was the national researcher who carried out the Kyrgyzstan assessment for the index, so my view on the result is from inside the work, not from a citation of it.

I am writing this on the day the index was published. The KG Labs / kglabs.org series that later went deeper into the Central Asian regional picture — Central Asia in the GIRAI 2023 Assessment, Kyrgyzstan in the GIRAI 2023 Assessment, and the five-country series alongside — drew on the same data this national researcher work produced. This post is the witness-voice version of how I read the country score on the day the index landed.

What the index measures

The Global Index on Responsible AI evaluates 19 thematic areas of responsible AI across three dimensions:

• Human Rights and AI — which civil and political rights the country’s AI ecosystem either protects or risks.
• Responsible AI Governance — what governance frameworks the country has put in place (or has not).
• National Responsible AI Capacities — what research, advisory, and civil-society infrastructure exists to support responsible AI.

Inside each thematic area, the index looks at three ecosystem components: government frameworks, government action, and non-state initiatives. That structure is more granular than other AI-readiness indices in the field. The Oxford Insights index, for example, asks «is the country technically and institutionally ready for AI?» The GIRAI asks a different question: «is the country’s actual AI activity, in 2023, responsible?» These are not the same question and they do not produce the same ranking.

As the national researcher for Kyrgyzstan, my work was to conduct extensive expert interviews, gather primary-source documentation, and produce the country’s evidence base for the assessment team to score against the rubric. That work covered around twenty stakeholders — across government, the academic sector, civil society, and the private technical community — over a roughly six-week intake window.

What 79th out of 138 actually says

There is a temptation, with a ranking, to read it as a single number that says either «good» or «bad.» That is not what this kind of index is for. The single most useful reading of the Kyrgyzstan result is dimensional, not aggregate:

• Kyrgyzstan ranks 1st in Central Asia overall — ahead of Kazakhstan, Uzbekistan, Tajikistan, and Turkmenistan on the GIRAI rubric.
• The country scores best on the Human Rights and AI dimension, where 8 of the 12 thematic areas turned up positive evidence.
• The country scores weakest on Responsible AI Governance — because the formal governance framework was, in 2023, still being drafted (the Digital Code’s Chapter 23 was the in-progress vehicle), and no adopted AI law existed.
• The country scores middle on National AI Capacities — Kyrgyz-language AI work is growing, the academic engagement is real, but the volume is small.

Translated out of the rubric, that picture is: Kyrgyzstan has the broadest human-rights footprint in Central Asia AI work, the thinnest formal governance framework, and a small-but-growing capacity base. The aggregate ranking of 79th is the geometric average of those three.

What the index does and does not see

The GIRAI is a snapshot. Its strength is that it forces an evidence-based answer to «what does responsible AI actually look like in this country in 2023.» Its weakness is that some of the most consequential work in Kyrgyzstan — the kind of community-network connectivity work, the Sanarip Insan digital-literacy programme, the early stages of the climate-AI collaboration that would later anchor in Istanbul — sits outside the formal AI-policy frame the index is structured around. That work shows up in the human-rights dimension where the rubric has a place for it. It does not show up where it should: under the broader question of «what kind of country-level AI ecosystem is forming here?»

That is not a critique of the index — every measurement tool draws a boundary somewhere. It is a note for the next reader. The GIRAI says Kyrgyzstan is 79th. What it does not say is whether Kyrgyzstan’s path to 50th or 30th runs through more formal governance documents, through more applied AI in public service, through bigger investments in non-state initiatives, or through some combination of all three. The answer to that is for the country’s policy community to argue out, with the GIRAI numbers as one input among several.

What I plan to do with the data

I intend to write country-level analytical briefs across Central Asia, the Caucasus, and Eastern Europe, drawing on the GIRAI evidence base plus the wider AI-policy and digital-rights literature, and to publish those briefs progressively over the next year. The Central Asian briefs landed first — that is what the kglabs.org GIRAI series in the AI governance category was — and the wider regional set will come through linkedin and through subsequent kglabs.org pieces.

If you are working on responsible AI policy in Central Asia or in an adjacent region, the GIRAI data is the strongest current evidence base to build from. The full source data, dimensional breakdowns, and methodological notes are at https://global-index.ai.

The view from the national researcher seat, on the day the index came out, is: the work is real, the gaps are visible, and the next year of policy work is what determines whether the score moves.