Artificial Intelligence in Education and the Digital Preconditions of Tertiary Expansion in Uzbekistan: ARDL Evidence from 2000–2023

Background: Direct annual national series on AI adoption in higher education are not consistently available for Uzbekistan, yet the diffusion of AI-enabled learning depends on measurable digital and economic preconditions. Methods: Using annual data for 2000–2023, this study models tertiary enrollment as a macro-level proxy for the expansion of AI-ready higher education, with internet use, mobile subscriptions, and real GDP per capita as explanatory factors in a trend-augmented ARDL/UECM framework. Trend-aware unit-root testing, lag selection, bounds testing, and residual diagnostics are implemented as one closed empirical sequence. Results: The preferred ARDL(1,3,1,1) specification supports cointegration, a significant error-correction mechanism, a positive long-run role for mobile access, and a negative internet coefficient after controlling for mobile inclusion, income, and structural trend. Conclusions: AI readiness in higher education should be interpreted as a conversion problem rather than a simple connectivity problem.