The essay examines how generative Artificial Intelligence is challenging the modern foundations of intellectual property. The author draws a historical parallel with medieval monastic culture, a period in which knowledge was regarded as a collective and sacred good, and creativity was understood as an act of compilation and humble imitation rather than individual originality. The modern concept of the “author” as the holder of exclusive rights over an economic asset emerged only later, with Humanism and the invention of the printing press. Today, the rise of general-purpose AI systems (GPAI) and large language models (LLMs) profoundly disrupts these modern legal paradigms. The essay highlights three main legal debates: the use of copyright-protected works as data for algorithmic training, the potentially derivative nature of AI outputs, and the possibility of granting legal protection to creations autonomously generated by AI. To address the structural tension between algorithmic innovation and the protection of pre-existing works, the author suggests that free and open-source licensing models may offer an alternative regulatory framework for a more ethical, participatory, and sustainable development of artificial intelligence technologies.
Il saggio esamina come l'Intelligenza Artificiale generativa metta in discussione le fondamenta moderne della proprietà intellettuale. L'autore traccia un parallelo storico con la cultura monastica medievale, un periodo in cui la conoscenza era considerata un bene collettivo e sacro e la creatività era vista come un atto di compilazione e umile imitazione piuttosto che come originalità individuale. Il concetto moderno di "autore" titolare di diritti esclusivi su un bene economico è emerso solo in seguito, con l'Umanesimo e l'invenzione della stampa. Oggi, l'avvento dei sistemi di IA per scopi generali (GPAI) e dei modelli linguistici di grandi dimensioni (LLM) sconvolge profondamente questi paradigmi giuridici moderni. Il saggio evidenzia tre principali dibattiti legali: l'uso di opere protette da copyright come dati di addestramento algoritmico, la natura potenzialmente derivata degli output dell'IA e la possibilità di concedere protezione legale alle creazioni generate autonomamente dall'IA. Per affrontare la tensione strutturale tra l'innovazione algoritmica e la tutela delle opere preesistenti, l'autore propone che i modelli di licenza liberi e open-source possano offrire un quadro normativo alternativo per uno sviluppo più etico, partecipativo e sostenibile delle tecnologie di intelligenza artificiale
Towards a Knowledge Ecology: Open Source as a Framework for AI
Giuseppe Sanseverino
2026-01-01
Abstract
The essay examines how generative Artificial Intelligence is challenging the modern foundations of intellectual property. The author draws a historical parallel with medieval monastic culture, a period in which knowledge was regarded as a collective and sacred good, and creativity was understood as an act of compilation and humble imitation rather than individual originality. The modern concept of the “author” as the holder of exclusive rights over an economic asset emerged only later, with Humanism and the invention of the printing press. Today, the rise of general-purpose AI systems (GPAI) and large language models (LLMs) profoundly disrupts these modern legal paradigms. The essay highlights three main legal debates: the use of copyright-protected works as data for algorithmic training, the potentially derivative nature of AI outputs, and the possibility of granting legal protection to creations autonomously generated by AI. To address the structural tension between algorithmic innovation and the protection of pre-existing works, the author suggests that free and open-source licensing models may offer an alternative regulatory framework for a more ethical, participatory, and sustainable development of artificial intelligence technologies.| File | Dimensione | Formato | |
|---|---|---|---|
|
SANSEVERINO-50_Chapter 1-Navigating.pdf
solo utenti autorizzati
Tipologia:
Versione Editoriale (PDF)
Licenza:
Copyright dell'editore
Dimensione
159.18 kB
Formato
Adobe PDF
|
159.18 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

