This dissertation explores the organizational implications of the adoption of learning algorithms within public organizations. Moving beyond a strictly technical focus, the research examines the reciprocal relationship between artificial intelligence (AI) and organizations, exploring how AI transforms these organizations while existing practices simultaneously shape AI adoption and use. Organizations are not a straightforward phenomenon. Therefore the study looks at organizations through different lenses: the individual level (focusing on professional autonomy and tacit knowledge), the organizational level (focusing on multi-actor collaboration and boundaries), and the contextual level (focusing on adaptation to external requirements). Through three exploratory empirical studies, the research highlights the practical realities of public AI adoption. It finds that while co-creation between machine learning developers and end-users can blur certain knowledge boundaries, deeply rooted boundaries around language and understanding persist, and new boundaries can unexpectedly emerge. Additionally, an interactive development process is shown to increase the explainability of AI models primarily through organizational learning rather than traditional technical transparency. Furthermore, the introduction of Generative AI is expected to function as a sparring partner for professionals. However, this role will increase the reliance on professionals' tacit knowledge and demand room for active experimentation. Ultimately, the dissertation concludes that public AI adoption introduces increased organizational complexity, the emergence of hybrid professionals and time investments without clear returns. To adopt AI succesfully public organizations must thoughtfully experiment with and embrace AI to secure its long-term benefits. The success of this AI-driven transformation remains fundamentally dependent on well-developed professionals that interact with each other and who can profesionally assess, utilize, and counterbalance these new technologies.
La presente tesi analizza le implicazioni organizzative dell’adozione di algoritmi di apprendimento all’interno delle organizzazioni pubbliche. Andando oltre un approccio strettamente tecnico, la ricerca esamina la relazione reciproca tra intelligenza artificiale (IA) e organizzazioni, analizzando come l’IA trasformi tali organizzazioni mentre, al contempo, le pratiche esistenti influenzano l’adozione e l’uso dell’IA. Le organizzazioni non sono un fenomeno semplice. Pertanto, lo studio analizza le organizzazioni attraverso diverse prospettive: il livello individuale (con particolare attenzione all’autonomia professionale e alla conoscenza tacita), il livello organizzativo (con particolare attenzione alla collaborazione tra più attori e ai confini organizzativi) e il livello contestuale (con particolare attenzione all’adattamento alle esigenze esterne). Attraverso tre studi empirici esplorativi, la ricerca mette in luce le realtà pratiche dell’adozione dell’IA nel settore pubblico. Essa rileva che, sebbene la co-creazione tra sviluppatori di apprendimento automatico e utenti finali possa rendere meno netti alcuni confini della conoscenza, persistono confini profondamente radicati relativi al linguaggio e alla comprensione, e possono emergere inaspettatamente nuovi confini. Inoltre, si dimostra che un processo di sviluppo interattivo aumenta la spiegabilità dei modelli di IA principalmente attraverso l’apprendimento organizzativo piuttosto che attraverso la tradizionale trasparenza tecnica. Inoltre, si prevede che l’introduzione dell’IA generativa funga da sparring partner per i professionisti. Tuttavia, questo ruolo aumenterà la dipendenza dalla conoscenza tacita dei professionisti e richiederà spazio per la sperimentazione attiva. In definitiva, la tesi conclude che l’adozione dell’IA da parte delle organizzazioni pubbliche comporta una maggiore complessità organizzativa, l’emergere di figure professionali ibride e investimenti di tempo senza chiari ritorni. Per adottare con successo l’IA, le organizzazioni pubbliche devono sperimentarla e abbracciarla in modo ponderato per garantirne i benefici a lungo termine. Il successo di questa trasformazione guidata dall’IA rimane fondamentalmente dipendente da professionisti ben preparati che interagiscano tra loro e che siano in grado di valutare, utilizzare e controbilanciare professionalmente queste nuove tecnologie.
Public organizations in transition: AI, professionals, and organizational change
van KRIMPEN, FLORIS JUSTUS
2026
Abstract
This dissertation explores the organizational implications of the adoption of learning algorithms within public organizations. Moving beyond a strictly technical focus, the research examines the reciprocal relationship between artificial intelligence (AI) and organizations, exploring how AI transforms these organizations while existing practices simultaneously shape AI adoption and use. Organizations are not a straightforward phenomenon. Therefore the study looks at organizations through different lenses: the individual level (focusing on professional autonomy and tacit knowledge), the organizational level (focusing on multi-actor collaboration and boundaries), and the contextual level (focusing on adaptation to external requirements). Through three exploratory empirical studies, the research highlights the practical realities of public AI adoption. It finds that while co-creation between machine learning developers and end-users can blur certain knowledge boundaries, deeply rooted boundaries around language and understanding persist, and new boundaries can unexpectedly emerge. Additionally, an interactive development process is shown to increase the explainability of AI models primarily through organizational learning rather than traditional technical transparency. Furthermore, the introduction of Generative AI is expected to function as a sparring partner for professionals. However, this role will increase the reliance on professionals' tacit knowledge and demand room for active experimentation. Ultimately, the dissertation concludes that public AI adoption introduces increased organizational complexity, the emergence of hybrid professionals and time investments without clear returns. To adopt AI succesfully public organizations must thoughtfully experiment with and embrace AI to secure its long-term benefits. The success of this AI-driven transformation remains fundamentally dependent on well-developed professionals that interact with each other and who can profesionally assess, utilize, and counterbalance these new technologies.| File | Dimensione | Formato | |
|---|---|---|---|
|
BiblioPDF Public Organizations in Transition_Floris van Krimpen.pdf
accesso aperto
Licenza:
Tutti i diritti riservati
Dimensione
3.2 MB
Formato
Adobe PDF
|
3.2 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/385230
URN:NBN:IT:POLIMI-385230