The Role of AI-Driven Automation Exposure in Shaping the Productivity Effects of European Intangible Capital Investments

This thesis researches the role of artificial intelligence-driven automation exposure in impacting the productivity effects of intangible capital investments in different European sectors. AI, recognized as general-purpose technology, has the potential to augment knowledge-intensive work while displacing routine tasks. The central research question is how sectoral differences in AI-driven automation potential influence the productivity returns from intangible capital. Using GLOBALINTO innovation survey, this study develops sector-specific investment multipliers based on predicted occupational shifts caused by AI-driven automation. These multipliers are applied to intangible variables to account for AI-impacted changes in nature of work. The analysis distinguishes between innovative (IC) and non-innovative (NOIC) occupations and models how AI alters their proportions, with the assumption that only a small portion of displaced NOIC workers can be translated into IC roles. Regression results show that organizational capital is consistently and positively associated with labor productivity, while AI exposure does not amplify this effect. R&D employment is negatively associated with productivity in short term, but the effect is slightly mitigated when AI exposure is considered, suggesting potential long-run complementarities. ICT investments do not show direct productivity gains but are strongly linked to innovation activity. These results highlight that intangible asset: organizational capital, ICT, and R&D function not in isolation, but synergistically. Their combined presence supports firm’s ability to adapt to AI-driven change. Moreover, AI increases the scaling effect of intangible assets by expanding intangibles and further increasing their collective impact on productivity and innovation. The thesis contributes to the literature by linking AI’s transformative potential with the productivity of intangible capital, emphasizing the need for sector-specific and task-level understanding. Limitations include the cross-sectional nature of the data, simplified assumptions in multiplier construction, and the challenge of fully capturing dynamic restructuring processes. Future research could build on this work by employing time-series data, firm-level AI adoption metrics, and improved measures of organizational transformation.