Armenia's Ministry of Labor Pilots Machine Learning to Target Social-Protection Beneficiaries

In October–November 2022, Armenia's Ministry of Labor and Social Affairs presented a machine-learning system built with the state-run Nork Technology Center and financed by the Asian Development Bank, intended to help the country's roughly 1.1 million social-protection beneficiaries. Ministry and ADB officials described it as the first application of AI to social-protection targeting anywhere in the world.

The model, built over one year on top of three years of prior research, was designed for five functions: checking benefit decisions against legal requirements; identifying beneficiaries likely to stay on welfare programmes beyond three years; identifying beneficiaries likely to exit due to rising income; clustering beneficiaries algorithmically to surface hidden patterns; and evaluating the effectiveness of state employment programmes. Nork Center director Anahit Parzyan described the system as autonomous, generating a new response for each new survey round; Labor Minister Narek Mkrtchyan called it a new policy approach 'not only in Armenia, but around the world'.

No independent evaluation, deployment scale-up, or outcome data has been published in the years since the 2022 presentation — repeated searches through September 2026 turn up only the original launch coverage. The practice is recorded here as a genuine, well-funded pilot with a clear methodology, but one whose real-world accuracy and impact on beneficiaries remain unverified and, on current public evidence, possibly stalled after the initial presentation.

Read the full analysis: https://caucasuswatch.de/en/news/armenia-becomes-first-in-world-to-apply-ai-in-social-protection.html