The thesis addresses the duty of directors of corporations to act in an informed manner when an artificial intelligence system is employed in support of managerial decision-making processes. Firstly, the paper examines the essential characteristics of machine learning and deep learning systems, as well as the nature of the outputs they generate. Against this backdrop, the main doctrinal theories concerning the integration of such systems within the scope of managerial functions are critically reviewed. The core of the research focuses on the use of AI systems as “investigatory tools”, with the aim of assessing their impact on intra-board information flows and outlining the conduct-related duties of directors who make use of them – drawing particular inspiration from the provisions of the AI Act. Indeed, board members are primarily required to comply with the general duty to act in an informed manner and to concretely adapt such duty in light of the decision-support tools. Finally, the thesis addresses liability profiles arising from breaches of the duty to act in an informed manner by directors using AI systems, also examining the additional remedies provided under the legal framework.
La tesi ha ad oggetto il dovere di agire in modo informato degli amministratori di società per azioni ove sia utilizzato un sistema d’intelligenza artificiale a supporto dei procedimenti decisionali gestori. In primo luogo, lo scritto esamina le caratteristiche essenziali dei sistemi di machine learning e di deep learning e la natura informativa degli output da questi prodotti. Ciò posto, vengono passate in rassegna, in chiave critica, le principali ipotesi formulate dalla dottrina chiamata ad inquadrare l’introduzione di tali sistemi nell’ambito della funzione gestoria. Il corpo centrale della ricerca approfondisce l’utilizzo del sistema di I.A. come strumento in funzione istruttoria, al fine di esaminarne l’impatto sui flussi informativi endoconsiliari e di delineare le regole di condotta degli amministratori che vi fanno ricorso, prendendo spunto, in particolare, dalle disposizioni dell’AI Act. I membri del plenum sono, infatti, chiamati a conformarsi al dovere autonomo a contenuto generale di cui all'art. 2381, c. 6, c.c. e a concretizzarlo individuando le subnorme di diritto positivo rilevanti in relazione al mezzo istruttorio utilizzato. Infine, il lavoro considera i profili di responsabilità per la violazione di tali subnorme, vagliando anche gli ulteriori rimedi approntati dall’ordinamento.
Il dovere di agire informato e il ruolo dell'intelligenza artificiale nei procedimenti decisionali degli amministratori di SpA. Le regole di condotta e la loro concretizzazione.
MORATO, COSTANZA
2026
Abstract
The thesis addresses the duty of directors of corporations to act in an informed manner when an artificial intelligence system is employed in support of managerial decision-making processes. Firstly, the paper examines the essential characteristics of machine learning and deep learning systems, as well as the nature of the outputs they generate. Against this backdrop, the main doctrinal theories concerning the integration of such systems within the scope of managerial functions are critically reviewed. The core of the research focuses on the use of AI systems as “investigatory tools”, with the aim of assessing their impact on intra-board information flows and outlining the conduct-related duties of directors who make use of them – drawing particular inspiration from the provisions of the AI Act. Indeed, board members are primarily required to comply with the general duty to act in an informed manner and to concretely adapt such duty in light of the decision-support tools. Finally, the thesis addresses liability profiles arising from breaches of the duty to act in an informed manner by directors using AI systems, also examining the additional remedies provided under the legal framework.| File | Dimensione | Formato | |
|---|---|---|---|
|
Morato_Tesi pdfA.pdf
embargo fino al 12/03/2027
Licenza:
Tutti i diritti riservati
Dimensione
3.97 MB
Formato
Adobe PDF
|
3.97 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/361630
URN:NBN:IT:UNIVE-361630