Conventus Law

Conventus Law

by

Introduction 

The EU AI Act defines Generative AI as “foundation models used in AI systems specifically intended to generate, with varying levels of autonomy, content such as complex text, images, audio, or video.” (Art. 28b (4) AI Act). Recent advancements, such as multi-modal systems capable of processing and generating different forms of content within a single framework, demonstrate that generative AI is no longer an experimental technology confined to specialist environments. Instead, it has become an accessible and widely used tool in everyday creative and commercial activities. 

This technological shift has renewed a fundamental legal question: to what extent does copyright law apply to works produced by, or with the assistance of, artificial intelligence? Traditional copyright doctrines of authorship and originality were developed on the assumption that creative works are the result of human intellectual effort. The growing autonomy of generative AI systems challenges this assumption and places pressure on existing legal frameworks. 

Unlike conventional software, these systems may operate with minimal human intervention, raising complex issues concerning creative control, ownership, and legal responsibility. This article focuses on generative AI outputs and examines how different legal systems distinguish between AI-assisted works and fully autonomous AI-generated content, particularly in relation to authorship and copyright protection. 

Different jurisdictions have responded to these challenges in markedly different ways. This article examines three contrasting approaches: Italy’s reaffirmation of human authorship, Ukraine’s introduction of a sui generis regime for AI-generated outputs, and China’s evolving judicial treatment of AI-assisted creativity. 

The Italian Approach: Reaffirming Human Authorship 

Italy is the first country to adopt a law on the protection of AI (Law 132/2025), which came into force on 10 October 2025. The legislation is based on and integrates EU Regulation 2024/1869, adopted on 13 June 2024. Italy has addressed the copyright implications of artificial intelligence through targeted amendments introduced by Article 25 of the Italian Artificial Intelligence Law. Section IV of the Italian AI Law contains only this provision, which makes two significant amendments to Law No. 633 of 1941, Italy’s longstanding Copyright Act (Legge sul Diritto d’Autore, “LDA”). These amendments are aimed at responding to the growing challenges posed by AI technologies while preserving the traditional foundations of copyright law. 

The recent law introduces amendments to Italian Copyright Law, whereby Article 1 now reads as follows: 

“Works of ‘human’ intellectual creation of a creative nature are protected under this law, including those belonging to literature, music, figurative arts, architecture, theatre, and cinematography, regardless of the mode or form of expression, even when created with the aid of artificial intelligence tools, provided they are the result of the author’s intellectual effort.” 

This amendment expressly confirms that copyright protection extends to works created with the assistance of AI systems, so long as the work remains the product of a creative and original human intellectual contribution. At the same time, it implicitly excludes fully autonomous AI-generated outputs from copyright protection, as such outputs lack the necessary human intellectual effort. 

Under this approach, AI may function as a creative tool, but copyright subsists only where a human author exercises meaningful creative control. Outputs generated entirely by AI systems without significant human involvement fall outside the scope of copyright protection. Italy therefore firmly rejects AI authorship and reinforces the traditional link between copyright and human intellectual effort. 

Ukraine’s Sui Generis Regime for AI-Generated Outputs 

In contrast to Italy’s human-centric approach, Ukraine has adopted a more innovative legislative solution. In 2022, Ukraine amended its Law on Copyright and Related Rights to introduce a sui generis form of protection specifically designed to regulate outputs generated without human involvement. 

The purpose of this reform is to address a legal gap by providing protection for AI-generated outputs that do not satisfy traditional copyright requirements. Under Ukrainian law, such outputs are classified as non-original objects generated by a computer program. For the protection of an AI-generated output, Art. 33(1) of the Ukrainian Copyright Law provides for two criteria: firstly, it is an object that differs from existing similar objects; secondly, it is formed as a result of the functioning of a computer program without the direct participation of an individual in the formation of this object. At the same time, works created by humans using computer technologies are expressly excluded from this category and remain protected under ordinary copyright law. 

Sui generis protection lasts 25 years from 1 January of the year following creation and covers only economic rights, as AI-generated outputs are not recognised as “works” and do not attract moral rights. Rights may vest in parties connected to the AI system, such as the initiator, developer, or licensee. Protection applies where the output is novel, automatically generated by software, and created without human creative input beyond activation. 

This regime ensures legal certainty for commercially valuable AI-generated content without redefining authorship, supplementing traditional copyright law with a tailored mechanism for automated creation. 

Judicial Approaches in China 

China on the other hand offers a contrasting model, relying primarily on judicial interpretation rather than legislative reform. Chinese courts have adopted a pragmatic, case-by-case approach that focuses on the extent of human intellectual contribution involved in the creation of AI-generated outputs. 

In Li Yunkai v. Liu Yuanchun (2023) Jing 0491 Min Chu No. 11279 (2023), the Beijing Internet Court held that an AI-generated image produced using Stable Diffusion qualified for copyright protection. The court applied four factors, assessing whether the output fell within the fields of literature, art, or science, possessed originality, was expressed in a tangible form, and resulted from intellectual achievement. 

The court held that the plaintiff’s active design choices through prompts, layout, composition, and iterative refinements demonstrated sufficient human creativity to confer originality. While AI systems and developers were not recognised as authors, the plaintiff was deemed the author due to his creative control over the image-generation process. This reasoning was later reinforced by the Changshu People’s Court in 2025, further confirming that users who exercise deliberate and creative control over AI outputs may qualify for authorship. Chinese courts have also emphasised transparency and good faith disclosure of AI usage, reflecting an approach that seeks to balance copyright protection with technological innovation. 

Conclusion 

Italy, Ukraine, and China adopt different approaches but share one principle: AI systems are not recognised as legal authors, and copyright protection depends on human intellectual contribution. Italy excludes fully autonomous AI outputs, Ukraine introduces a sui generis regime for non-original AI content, and China recognises copyright in AI-assisted works where users exercise meaningful creative control. 

These divergent approaches reflect the lack of international harmonisation and differing policy priorities. As generative AI becomes more autonomous and widespread, copyright law will need clearer standards and greater transparency to remain effective and relevant. 

For further information, please contact:

Ahmad Hafiz Zubir, Partner, Azmi & Associates

hafiz.zubir@azmilaw.com

References: 

Register for your monthly Asia legal updates from Conventus Law

Error: Contact form not found.

RELATED ARTICLES