This thesis investigates the relationship between artificial intelligence (AI) regulation and innovative entrepreneurship, focusing on the European Union Artificial Intelligence Act (AI Act). While AI has become a central driver of economic and technological transformation, its rapid expansion has prompted growing regulatory attention. Against this background, the study tests the hypothesis that the announcement of the EU AI Act negatively affected AI startup creation within the European Union compared to other regions. The analysis follows a mixed-method approach. The qualitative component examines AI policy frameworks across major global regions using NVivo and large language model–based text analysis. The findings reveal significant differences in regulatory orientation: China exhibits a restrictive leaning and security-driven regulation; the European Union combines innovation promotion with risk and ethics-oriented governance, leaning restrictive; and the United States maintains a more liberal, innovation-focused narrative. The quantitative analysis employs a Difference-in-Differences econometric model using Crunchbase data on over 55,000 AI startups from 2005 to 2023, comparing trends in the EU and the United States before and after the AI Act announcement in April 2021. Results indicate a significant and robust decline in AI startup formation in the EU relative to the US, consistent across multiple robustness checks and additional analyses that took in consideration the rest of the world and the effects of the General Data Protection Regulation (GDPR). Overall, the study demonstrates that regulatory uncertainty and anticipated compliance costs can temporarily discourage entrepreneurial activity, even when regulation aims to promote trustworthy innovation. The results underscore the importance of regulatory predictability, early communication, and accessible innovation sandboxes as policy tools to reconcile technological safety with sustained entrepreneurial growth.
Questa tesi indaga la relazione tra la regolamentazione dell’intelligenza artificiale (AI) e l’imprenditorialità innovativa, concentrandosi in particolare sull’Artificial Intelligence Act (AI Act) dell’Unione Europea. Sebbene l’AI sia divenuta un motore centrale della trasformazione economica e tecnologica, la sua rapida espansione ha suscitato un crescente interesse regolatorio. In questo contesto, lo studio verifica l’ipotesi secondo cui l’annuncio dell’AI Act europeo abbia influenzato negativamente la creazione di startup nel settore dell’AI all’interno dell’Unione Europea rispetto ad altre regioni. L’analisi adotta un approccio a metodi misti. La componente qualitativa esamina i quadri normativi in materia di AI nelle principali aree globali utilizzando il software NVivo e un’analisi testuale basata su Large Language Model (LLM). I risultati mostrano differenze significative negli orientamenti regolatori: la Cina presenta un’impostazione restrittiva e orientata alla sicurezza; l’Unione Europea combina la promozione dell’innovazione con una governance centrata su rischio ed etica, tendendo anch’essa verso un approccio restrittivo; mentre gli Stati Uniti mantengono un orientamento più liberale e focalizzato sull’innovazione. L’analisi quantitativa impiega un modello econometrico Difference-in-Differences basato su dati di oltre 55.000 startup AI raccolti da Crunchbase per il periodo 2005–2023, confrontando l’andamento della nascita di startup in Europa e negli Stati Uniti prima e dopo l’annuncio dell’AI Act (aprile 2021). I risultati evidenziano un calo significativo e robusto della formazione di nuove startup AI nell’UE rispetto agli Stati Uniti, confermato da numerosi robustness checks e analisi aggiuntive. Al contrario, l’entrata in vigore del Regolamento Generale sulla Protezione dei Dati (GDPR) nel 2018 ha mostrato un effetto solo lieve e temporaneo. Nel complesso, lo studio dimostra che l’incertezza normativa e i costi di conformità previsti possono temporaneamente scoraggiare l’attività imprenditoriale, anche quando la regolazione mira a promuovere un’innovazione affidabile. I risultati sottolineano l’importanza della prevedibilità normativa, della comunicazione tempestiva e dell’accesso a sandbox regolatorie come strumenti chiave per conciliare la sicurezza tecnologica con una crescita imprenditoriale sostenuta.
The impact of AI regulation on startup formation: evidence from European Union AI Act
Napodano, Matteo;SERATONI GUALDONI, GIACOMO
2024/2025
Abstract
This thesis investigates the relationship between artificial intelligence (AI) regulation and innovative entrepreneurship, focusing on the European Union Artificial Intelligence Act (AI Act). While AI has become a central driver of economic and technological transformation, its rapid expansion has prompted growing regulatory attention. Against this background, the study tests the hypothesis that the announcement of the EU AI Act negatively affected AI startup creation within the European Union compared to other regions. The analysis follows a mixed-method approach. The qualitative component examines AI policy frameworks across major global regions using NVivo and large language model–based text analysis. The findings reveal significant differences in regulatory orientation: China exhibits a restrictive leaning and security-driven regulation; the European Union combines innovation promotion with risk and ethics-oriented governance, leaning restrictive; and the United States maintains a more liberal, innovation-focused narrative. The quantitative analysis employs a Difference-in-Differences econometric model using Crunchbase data on over 55,000 AI startups from 2005 to 2023, comparing trends in the EU and the United States before and after the AI Act announcement in April 2021. Results indicate a significant and robust decline in AI startup formation in the EU relative to the US, consistent across multiple robustness checks and additional analyses that took in consideration the rest of the world and the effects of the General Data Protection Regulation (GDPR). Overall, the study demonstrates that regulatory uncertainty and anticipated compliance costs can temporarily discourage entrepreneurial activity, even when regulation aims to promote trustworthy innovation. The results underscore the importance of regulatory predictability, early communication, and accessible innovation sandboxes as policy tools to reconcile technological safety with sustained entrepreneurial growth.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_12_Napodano_Seratoni_Tesi.pdf
accessibile in internet per tutti
Descrizione: Tesi
Dimensione
10.76 MB
Formato
Adobe PDF
|
10.76 MB | Adobe PDF | Visualizza/Apri |
|
2025_12_Napodano_Seratoni_Executive Summary.pdf
accessibile in internet per tutti
Descrizione: Executive Summary
Dimensione
2.31 MB
Formato
Adobe PDF
|
2.31 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/246783