This thesis investigates the role of Artificial Intelligence in the transformation of contemporary Corporate Governance models, starting from the premise that AI is no longer merely a technological tool aimed at improving the efficiency of business processes, but is progressively emerging as a new paradigm capable of affecting the ways in which decision-making power is exercised, as well as corporate control and accountability mechanisms. The integration of algorithms into strategic and operational decision-making raises novel questions: who is responsible for a decision supported by AI? How are directors’ fiduciary duties changing? What skills should corporate leaders possess, and what mechanisms can be used to protect shareholders? To address these questions, the research adopts a comparative approach and examines three global technology companies, Alphabet, Microsoft, and IBM, analysing their strategies across four key dimensions: the governance of AI-related risks, the evolution of fiduciary duties, the changing competencies required of Boards of Directors, and the management of agency conflicts. The analysis highlights the emergence of a shared governance model in which algorithmic innovation is accompanied by new structures of oversight and accountability. All three organisations strengthen oversight bodies and committees, introduce structured procedures for technical assessment, and adopt mechanisms designed to ensure that AI outputs are interpretable and understandable to those responsible for making decisions. At the same time, a significant shift is emerging in the composition of competencies at the highest levels of corporate leadership: knowledge of AI is no longer regarded as a skill confined to a limited number of specialists, but is becoming a strategic competence increasingly embedded across corporate leadership. From this perspective, the central challenge posed by AI to corporate governance lies not merely in governing a new technology, but in rethinking the relationship between algorithms, decision-making power, and accountability. Such a rethinking, in turn, calls for a broader conception of corporate responsibility, one that cannot be confined to the strictly internal dimension, but must also take into account the environmental, social, and governance impacts arising from the growing integration of AI and digital technologies into business activities. The analysis therefore shows that leading companies are moving from a logic of simply adopting AI towards a logic of AI governance, in which transparency, oversight, competence, and accountability are not only necessary conditions for turning technological innovation into sustainable value for the company and its stakeholders, but also constitute the foundations of a broader form of Corporate Digital Responsibility. This points to a model in which governing AI means not only ensuring its effective and compliant use, but also anticipating, assessing, and assuming responsibility for its impacts, thereby strengthening the link between technological innovation, corporate governance, and corporate responsibility.
La presente tesi si propone di indagare il ruolo dell’Intelligenza Artificiale nella trasformazione dei modelli contemporanei di Corporate Governance, muovendo dalla considerazione che l’IA non costituisca più un mero strumento tecnologico funzionale all’efficientamento dei processi aziendali, ma si configuri progressivamente come un nuovo paradigma in grado di incidere sulle modalità di esercizio del potere decisionale, sui sistemi di controllo e sui meccanismi di responsabilità societaria. L’ingresso degli algoritmi nelle scelte strategiche e operative solleva infatti interrogativi inediti: chi è responsabile di una decisione supportata dall’IA? Come cambiano i doveri fiduciari degli amministratori? Quali competenze devono possedere i vertici aziendali e quali strumenti possono tutelare gli azionisti? Per rispondere a queste domande, la ricerca adotta un approccio comparativo e prende in esame tre protagonisti globali del settore tecnologico, Alphabet, Microsoft e IBM, analizzandone le strategie attraverso quattro dimensioni fondamentali: la gestione dei rischi legati all’IA, l’evoluzione dei doveri fiduciari, la trasformazione delle competenze richieste ai Consigli di Amministrazione e la gestione dei conflitti d’agenzia. L’analisi evidenzia l’emergere di un modello di governance condiviso, nel quale l’innovazione algoritmica viene accompagnata da nuove architetture di controllo e responsabilità. Le tre organizzazioni rafforzano gli organi e i comitati di supervisione, introducono procedure strutturate di verifica tecnica e adottano meccanismi volti a garantire che gli output dell’IA siano interpretabili e comprensibili da parte di chi è chiamato a prendere le decisioni. Parallelamente, emerge un cambiamento significativo nella composizione delle competenze ai vertici: la conoscenza dell’IA non viene più considerata una prerogativa riservata a pochi specialisti, ma diventa una competenza strategica diffusa all’interno della leadership aziendale. In questa prospettiva, la vera sfida dell’IA per la corporate governance non consiste soltanto nel governare una nuova tecnologia, ma nel ripensare il rapporto tra algoritmi, potere decisionale e responsabilità. Un simile ripensamento implica, a sua volta, una concezione più ampia della responsabilità aziendale, che non può essere circoscritta alla dimensione strettamente interna, ma deve necessariamente tenere conto degli impatti ambientali, sociali e di governance derivanti dalla crescente integrazione dell’IA e delle tecnologie digitali nelle attività d’impresa. L’analisi mostra quindi come le imprese più avanzate stiano passando da una logica di semplice adozione dell’IA a una logica di governance dell’IA, nella quale trasparenza, controllo, competenza e accountability non rappresentano soltanto condizioni necessarie per trasformare l’innovazione tecnologica in valore sostenibile per l’impresa e per i suoi stakeholder, ma costituiscono anche i presupposti di una più ampia responsabilità digitale dell’impresa. Ne emerge un modello nel quale governare l’IA significa non soltanto assicurarne un impiego efficace e conforme, ma anche anticiparne, valutarne e assumere responsabilmente gli impatti, rafforzando il legame tra innovazione tecnologica, corporate governance e responsabilità d’impresa.
Ripensare la Corporate Governance nell’Era dell’Intelligenza Artificiale: un’Analisi Comparativa di Alphabet, Microsoft e IBM
TODISCO, ALESSANDRA
2025/2026
Abstract
This thesis investigates the role of Artificial Intelligence in the transformation of contemporary Corporate Governance models, starting from the premise that AI is no longer merely a technological tool aimed at improving the efficiency of business processes, but is progressively emerging as a new paradigm capable of affecting the ways in which decision-making power is exercised, as well as corporate control and accountability mechanisms. The integration of algorithms into strategic and operational decision-making raises novel questions: who is responsible for a decision supported by AI? How are directors’ fiduciary duties changing? What skills should corporate leaders possess, and what mechanisms can be used to protect shareholders? To address these questions, the research adopts a comparative approach and examines three global technology companies, Alphabet, Microsoft, and IBM, analysing their strategies across four key dimensions: the governance of AI-related risks, the evolution of fiduciary duties, the changing competencies required of Boards of Directors, and the management of agency conflicts. The analysis highlights the emergence of a shared governance model in which algorithmic innovation is accompanied by new structures of oversight and accountability. All three organisations strengthen oversight bodies and committees, introduce structured procedures for technical assessment, and adopt mechanisms designed to ensure that AI outputs are interpretable and understandable to those responsible for making decisions. At the same time, a significant shift is emerging in the composition of competencies at the highest levels of corporate leadership: knowledge of AI is no longer regarded as a skill confined to a limited number of specialists, but is becoming a strategic competence increasingly embedded across corporate leadership. From this perspective, the central challenge posed by AI to corporate governance lies not merely in governing a new technology, but in rethinking the relationship between algorithms, decision-making power, and accountability. Such a rethinking, in turn, calls for a broader conception of corporate responsibility, one that cannot be confined to the strictly internal dimension, but must also take into account the environmental, social, and governance impacts arising from the growing integration of AI and digital technologies into business activities. The analysis therefore shows that leading companies are moving from a logic of simply adopting AI towards a logic of AI governance, in which transparency, oversight, competence, and accountability are not only necessary conditions for turning technological innovation into sustainable value for the company and its stakeholders, but also constitute the foundations of a broader form of Corporate Digital Responsibility. This points to a model in which governing AI means not only ensuring its effective and compliant use, but also anticipating, assessing, and assuming responsibility for its impacts, thereby strengthening the link between technological innovation, corporate governance, and corporate responsibility.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi magistrale .pdf
non disponibili
Descrizione: La tesi analizza il ruolo dell’Intelligenza Artificiale nella trasformazione della corporate governance, con un confronto tra Alphabet, Microsoft e IBM.
Dimensione
2.98 MB
Formato
Adobe PDF
|
2.98 MB | Adobe PDF | Richiedi una copia |
È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: [email protected].
https://hdl.handle.net/20.500.14239/36832