Digital Copyright

The conflict over AI-produced art and music is often framed as a new copyright problem, but its deepest source is a crisis in cultural intuition. Human beings have always learned by imitation, quotation, influence, parody, apprenticeship, variation, and style transfer. Musicians learn from earlier musicians. Painters absorb composition, gesture, color, and technique. Quilters repeat inherited patterns. Mechanics teach one another how to repair carburetors. Communities share methods for cooking, building, mending, storytelling, singing, and surviving. Culture depends on this circulation. Yet many of the same people who tolerate human imitation react with moral alarm when an AI system produces an image, song, or voice "in the style of" a living artist. This paper argues that the alarm is not merely inconsistency. AI changes the social meaning of imitation by making it scalable, automated, opaque, commercial, and detached from ordinary human apprenticeship. The resulting turmoil arises because AI collapses the familiar category of cultural participation into the unfamiliar category of industrial extraction.

The paper proposes a Cultural Balance Model for evaluating AI art and music. The model weighs public benefit, transformative contribution, consent or compensation, and transparency against market harm, identity appropriation, scale, and deception. It is not offered as a mechanical substitute for judgment, but as a discipline for judgment: a way to separate disagreements about weights from disagreements about facts. The central thesis is that humans tolerate imitation when it feels like participation in culture, but resist it when it feels like extraction from culture. A stable settlement in the AI age must therefore preserve the cultural commons while protecting persons from being turned into unconsenting inputs, substitutes, or digital masks. The guiding principle is this: culture must remain learnable, but persons must not become extractable.

This chapter examines the scope of copyright protection for architectural works, with particular
emphasis on the legal challenges surrounding the definition of a “building” and the protection
of non-habitable structures. It explores the historical development of copyright protection for
architectural works and highlights the uncertainties that arise when distinguishing protected
architectural works from non-protected three-dimensional structures such as bridges, dams, and
monuments. The study compares the legal frameworks governing architectural copyright in the
United Kingdom and Zambia. It finds that while both jurisdictions recognize architectural
works as copyrightable artistic works, the United Kingdom provides a broader and clearer
definition of a building under its copyright legislation, thereby offering greater legal certainty.
In contrast, the Zambian Copyright Act does not define the term “building,” creating ambiguity
regarding the scope of protection available to architects and designers. The chapter further
analyzes the remedies available for copyright infringement, including damages, injunctions,
accounts of profits, delivery up of infringing articles, and criminal sanctions. It observes that
although both jurisdictions provide similar remedies and comply with international copyright
standards, the absence of clear statutory definitions and judicial guidance in Zambia limits the
effective enforcement of rights relating to architectural works. The study concludes that legal
clarity regarding the protection of architectural works is essential to safeguard the rights and
interests of creators while encouraging innovation and development within the architectural
profession.

Il recente caso “GEMA” del 11 novembre 2025 cattura l’attenzione perché chiarisce il funzionamento dei trasformatori generativi pre-addestrati (GPT) e, più in generale, le modalità di addestramento di un modello LLM, offrendo allo stesso tempo spunti interessanti per bypassare le difficoltà tecniche dovute allo scraping, ossia all’estrazione dei dati dai siti web effettuata attraverso specifici software

Resumen.-Este estudio examina como la IA generativa afecta el derecho de autoria en el Perú, la investigación se centra en dos cuestiones principales, el desarrollo de creaciones protegidas y tambien de contenidos en las redes sociales. La generación de contenidos utilizando IA en nuestra legislación peruana crea vacíos legales de reconocimiento de autoría y responsabilidad legal por obras generadas algorítmicamente. Además, encontró que entrenar modelos de IA puede afectar los derechos de reproducción y modificación de los autores, mientras que el contenido generado por IA plantea dudas sobre su protección legal debido a la falta de intervención humana directa. El trabajo sigue un enfoque cualitativo, dogmático jurídico y hermenéutico basado en análisis doctrinal, normativo y jurisprudencial. Se concluye que es un imperativo actualizar las leyes peruanas para alinear el normal desarrollo y creación progresivo de información e inovacines en avances tecnologicos para que el creador de datos informaticos tenga protección de autoria en la IA generativa en el Perú.