AI at Work: Promises, Perils and Paradoxes of Adoption

	PDF PHD_Dissertation_final.pdf Download (6MB)
	PDF thesis_booklet.pdf Download (103kB)
	PDF PHD_Dissertation_final.Text.Marked.pdf Download (6MB)

Abstract

This dissertation examines the complex labour market implications of artificial intelligence (AI) adoption, focusing on the Central and Eastern European context. Treating AI as a general-purpose technology, it analyses how automation and augmentation dynamics reshape employment, skill demands, and organisational adaptation. The study integrates bibliometric, empirical, and survey-based approaches to identify key mechanisms and policy challenges of AI-driven transformation. A scientometric meta-analysis of 250 peer-reviewed articles revealed persistent “cautious pessimism” in the literature: most works expect job displacement to outweigh creation, though skill-biased complementarities emerge in some sectors. Interviews with Central European professionals, triangulated with global developer data, confirmed this duality—senior employees reported productivity gains, while junior staff feared displacement, creating a “junior paradox.” A large-scale NLP analysis of 1.3 million Hungarian job advertisements quantified AI exposure across industries, showing concentrated adoption in ICT and manufacturing, but limited diffusion elsewhere. An international organisational survey further identified trust, governance, and quality assurance as critical enablers of successful AI integration. Results demonstrate that AI simultaneously creates and destroys jobs, amplifies wage and regional disparities, and shifts competitive advantage towards high-skill, AI-complementary roles. Unequal adoption patterns, weak institutional trust, and inadequate reskilling frameworks hinder inclusive growth. The dissertation concludes that sustainable technological transition requires coordinated policy measures promoting transparency, workforce development, and trust governance to ensure equitable and resilient adaptation to the AI economy.

View Item