Top 50%
dans ce catalogue (535 pratiques notées)
Les scores se concentrent dans le haut de l’échelle ; la position au sein du catalogue est donc souvent plus parlante que le chiffre seul.
Evidence of impact / measured public value2/3
Concrete, government-verified figures are public - 368 of 2,000+ lakes classified hazardous, 1,663m of diversion channels reinforced, 150 households directly protected - but no independent evaluation isolates the AI module's own contribution to warning accuracy or damage avoided.
Transparency, fairness & accountability1/3
UNDP and the Ministry of Emergency Situations publish programme updates and figures, but the AI/machine-learning module's methodology, accuracy metrics and thresholds are not published for independent review.
Transferability / demonstrated replication2/3
Tajikistan, Uzbekistan, Indonesia and Nepal have formally expressed interest in adopting the same monitoring approach, though none has yet implemented it, so replication is demonstrated as interest rather than completed transfer.
Scalability beyond pilot2/3
The system already operates nationally across all 368 classified hazardous lakes rather than a single-site pilot, integrated into a national 'Unified System of Comprehensive Monitoring and Forecasting'.
Governance, capability & sustainability2/3
Ownership sits with a named government ministry under a signed bilateral (Japan) and multilateral (UNDP) financing agreement, though continued funding after the current project phase is not yet confirmed in public sources.
Evidoria. Système unifié de surveillance par IA des inondations dues à la rupture de lacs glaciaires — le ministère des Situations d'urgence du Kirghizstan. ID persistant: 1941bbb5-6dc0-4ece-859a-7d9b335a7f17.