Regional Patterns and Disparities in Enterprise AI Adoption: Evidence from Hungary and Neighboring Countries

Šošić, Hana (2026) Regional Patterns and Disparities in Enterprise AI Adoption: Evidence from Hungary and Neighboring Countries. MA/MSc, Lámfalussy Sándor Közgazdaságtudományi Kar (volt Közgazdaságtudományi Kar).

PDF LKK-A-2026-77.pdf Hozzáférés joga: Csak nyilvántartásba vett egyetemi IP-kről nyitható meg Download (1MB)

Absztrakt (kivonat)

This thesis examines the structural determinants of enterprise artificial intelligence adoption across 27 NUTS-2 regions in Hungary, Austria, Croatia, Slovenia, and Slovakia, using Eurostat enterprise survey data for 2023–2025. The central research question asks why AI adoption rates vary so dramatically across regions in this part of Central Europe, from below 6 percent in the least digitally active areas to above 37 percent in the most advanced, and which regional characteristics best explain that variation. The study adopts a cross-sectional quantitative research design, estimating a series of OLS regression models supplemented by a fractional logit estimator and a pooled cross-sectional framework that exploits three annual observation waves. The core model regresses the regional share of enterprises using AI on three theoretically motivated predictors: GDP per capita in purchasing power standards, tertiary educational attainment, and high-technology employment as a share of total employment. The central finding is clear and consistent across every specification tested: regional GDP per capita is the dominant predictor of enterprise AI adoption. The relationship is large in magnitude, positive in sign, and statistically significant at the one percent level in all six model specifications. It holds when capital city regions are excluded, when the income variable is specified in levels rather than logarithms, under a non-linear estimator, and in a pooled model with year fixed effects and region-clustered standard errors. Tertiary educational attainment is strongly correlated with adoption in bivariate data but loses statistical significance once income is controlled, reflecting near-perfect collinearity between the two variables in this regional context rather than a substantive zero effect. High-technology employment shows a negative and sometimes significant coefficient in multivariate models, a result interpreted as reflecting the manufacturing-intensive character of CEE technology employment, which generates different spillovers than the software-intensive technology activity more prevalent in Western Europe. The pooled model reveals that average adoption roughly doubled across the sample between 2023 and 2025, yet the income gradient remained stable across all three waves. This pattern, where AI adoption is increasing but regional differences remain, suggests that AI tends to reinforce existing regional strengths rather than reduce them. This also raises concerns about the EU Digital Decade Program’s target of 75% enterprise AI adoption by 2030. The policy implication is that digital interventions operating exclusively at the technological level, AI literacy programs, innovation vouchers, cloud subsidies, are unlikely to close the adoption gap in structurally lagging regions if the underlying economic conditions remain unchanged. The CEE adoption gap is, at its core, a regional development problem that happens to be visible in digital statistics.

Magyar cím

Regionális mintázatok és különbségek a vállalati mesterségesintelligencia-alkalmazásban: bizonyítékok Magyarországról és a szomszédos országokból

Angol cím

Regional Patterns and Disparities in Enterprise AI Adoption: Evidence from Hungary and Neighboring Countries

Intézmény

Soproni Egyetem

Kar

Lámfalussy Sándor Közgazdaságtudományi Kar (volt Közgazdaságtudományi Kar)

Tanszékcsoport/intézet

LKK - Közgazdasági és Nemzetközi Kapcsolatok Intézet

Szak

LKK - Nemzetközi gazdaság és gazdálkodás mesterszak (MA)

Témavezető(k)

Helyi kari azonosító

LKK-A-2026-77

Mű típusa:	Diplomadolgozat (MA/MSc)
Felhasználói azonosító szám (ID):	Hana Sosic
Dátum:	21 Júl 2026 13:37
Utolsó módosítás:	21 Júl 2026 13:37
URI:	http://diploma.uni-sopron.hu/id/eprint/16770

Actions (login required)

Tétel nézet