Comparative Study and Insights on Copyright Protection of AIGC

from an International Perspective

Qi Jiang

Faculty of Law, The University of Sydney, Sydney 2006, Australia

Keywords: AIGC, Copyright Protection, Intellectual Property Rights, International Comparison.

Abstract: As artificial intelligence technology develops rapidly, Artificial Intelligence-Generated Content (AIGC) has

gradually permeated literature, art, and music, and changed the method of content creation and the landscape

of the traditional cultural industry. However, the lightning-fast proliferation of AIGC has brought about many

legal issues related to copyright protection. Based on an international viewpoint, this paper explores AIGC

paper argues for the reforms of the standards of originality, reformulating diversified copyright attribution

rules, and strengthening international cooperation and coordination, in order to build a copyright protection

framework suit for AIGC. Hoping to offer a glimpse and guidance for China and other countries to adjust

their AIGC copyright legal systems to promote the harmonious development of technology and law to en-

courage the innovation and sustainable development of cultural and creative industries.

1 INTRODUCTION

1.1 Research Background

With the speed of development of artificial

intelligence technology, from literature, art,

journalism, advertising, music, and film to more

fields, Artificial Intelligence-Generated Content

(AIGC) is gradually infiltrating everything, deeply

changing content creation methods and traditional

cultural industry patterns. AI-generated text, images,

and music are increasingly applied in commerce and

daily life. For example, ChatGPT is widely used in

content creation and customer service, while models

like DALL-E and MidJourney are applied in cover art

design, advertising, and entertainment (Li et al.,

2024).

But the swift proliferation of AI-generated

content has also accompanied many legal problems,

with disputes over their copyright protection

occupying the forefront of academia and industry.

Most of the current system of traditional copyright

protection is built on the basis of factors like human

creators creating ideas in an original manner and

personally having rights over them. This framework

fails against the characteristics of AIGC works which

lack a specific "human author", do not follow the

regular "creation process" (the "algorithms", big data,

training models, etc.), and therefore make it difficult

to determine their "originality" and "authorship"(Cai

et al., 2024).

1.2 Research Subject and Significance

The purpose of this study is to address the copyright

protection issues of AI-generated content, with a

special focus on the legal issues related to originality

determination and it is respective patent. AI-

generated content has brought forth new types of

creation and led to a lot of debates and controversies

in the copyright field. Thus, this research focuses on

studying the legislation and practices of various

countries in copyright protection of AIGC, whether

and how AIGC can be combined with the current

international copyright protection system.

The theoretical significance lies in addressing

these issues, and their resolution will have an

enormous impact on policy-making, shaping the

creation of content that will influence the future

development of the content creation industry. The

clarification of AIGC copyright attribution

encourages innovative applications of AI while

safeguarding the aforementioned legitimate rights of

content creators. This study aims to provide legal

references for China and other countries protecting

copyrights in the AIGC era, promoting the mutual

Jiang, Q.

Comparative Study and Insights on Copyright Protection of AIGC from an International Perspective.

DOI: 10.5220/0013962900004912

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Innovative Education and Social Development (IESD 2025), pages 54-59

ISBN: 978-989-758-779-5

development of technology and law, and continuing

to innovate and grow healthily within cultural and

creative industries.

2 CHARACTERISTICS OF AI-

GENERATED CONTENT AND

2.1 Technical Characteristics of

AI-Generated Content

AI-generated content is realized by algorithms, big

data, and deep learning technologies which allow AI

to work on a large amount of data within a short time

delivering the content in various s for-mats like text,

images, and music. As a result, AIGC possesses

special technical properties, i.e. massive data

processing, creative content generation, cross-modal

integration, and a certain extent of cognitive

interaction capability (Li et al., 2023). The creation

process of AI-generated content typically involves

several key steps:

• Data Training: AI models require extensive

data for training, which can include text,

images, audio, etc.. The richness and creativity

of AI-generated content are dependent on the

quality and quantity of training data they use.

•

Model Generation: AI models use deep

learning algorithms like Generative

Adversarial Net-works (GANs) or Natural

Language Processing models like

Transformers, to learn patterns from input

data, and generate outputs (China Academy of

Information and Communications

Technology, 2022).

•

User Input and Algorithm Output: Specific

instructions or parameters are provided by the

users and AI computes to deliver content that

matches the requirements (China Academy of

Information and Communications

Technology, 2022). This process differs from

conventional human creation, as it relies on re-

expression through data and algorithms,

without subjective creativity or emotional

input.

The creation characteristics of AIGC fundamentally

differ from human creative processes:

• Impersonality: Because of its emptiness of

subjective intent and emotional expression by

the author, AI-generated content lacks the

personalized elements characteristic of human

creation.

• Imitative Nature Based on Training Data: At its

core, AI-generated content is an imitation and

a recombination of its training data. Its

"creativity" is largely in an innovative and

reassembly of data already available, and not

the creation of totally new ideas (China

Academy of Information and

Communications Technology, 2022).

2.2 Core Issues in Copyright Disputes

AI-generated content sparks multiple disputes in the

realm of copyright protection, primarily focusing on

the following aspects:

2.2.1 Originality Determination

In copyright law, originality is one of its core

requirements. Traditional copyright law requires that

works be independently created by human authors

and that they have at least some creative expression.

However, this requirement presents huge challenges

to AIGC. The extent to which AI-generated content

meets the originality requirement is contentious

because AI-generated content is produced by learning

and imitating existing data (Wu, 2024).

However, some believe that while AI output can

be considered novel, it essentially consists of

randomizing a pre-existing arrangement of data, all

the while trying to imitate, and therefore infringing on

the human thoughtful effort and skills necessary to

drive traditional copyright law. As a result, it lacks

originality on the legal level (Yu, 2024). Still, others

argue that in the case of some AI-generated works,

their expression forms creative works as new and

original as to merit copyright protection.

2.2.2 Authorship and Copyright Ownership

A second important issue is the identification of the

"author," a principal concept of copyright law. Given

that AI is not a natural person and does not have legal

personality, determining the ownership of AI-

generated content is challenging. There are three

predominant views on this issue:

Attribution to AI Developers: Some argue that the

that created and trained the AI system since it

provides the technological scaffolding (or basis) for

the exploitation of content (Abbott, 2018).

Comparative Study and Insights on Copyright Protection of AIGC from an International Perspective

Attribution to Users: Others argue that those who

type a particular instruction and dictate creative

direction to the content generating process are to be

treated as authors, because those are critical

contributions (Wang, 2014).

Public Domain: A third view suggests that AI-

generated content should enter the public domain

because it does not meet the traditional definition of

authorship, and no human creation is directly

involved. In its Compendium of U.S. Copyright

Office Practices (Third Edition), the U.S. Copyright

Office explicitly states that works that do not have

contributions from human authors cannot be

registered as copyrighted. For example, the

guidelines make it clear that "works generated solely

by machines independently" cannot be copyrighted

(U.S. Copyright Office, 2021).

3 INTERNATIONAL COMPARI-

SON: LEGAL ATTITUDES

TOWARD COPYRIGHT

PROTECTION OF AIGC

3.1 The Legal Stance and Practice in

the United States

The key to copyright protection in the United States

is the recognition of a “human author”. The U.S.

beings (U.S. Copyright Office, 2021). In the Thaler

Case, however, this stance was explicitly confirmed

(U.S. Copyright Office Review Board, 2022). The

U.S. Copyright Office turned down an application for

a copyright on an AI-generated image in the Thaler

Case, as the work was deemed to be lacking human

authorship and originality. The case involved an AI

system developed by Stephen Thaler that produced an

image titled A Recent Entrance to Paradise. The core

reason for the rejection was the absence of any human

creative input in the work.

In other cases, like the Zarya of the Dawn case in

2022, the U.S. courts held that human arrangement

and editing could be protected, but AI-generated

elements without human inputs could not be

copyrighted (U.S. Copyright Office, 2023). These

cases demonstrate that the United States enforces a

high threshold for 'human creativity' in determining

involvement.

The U.S. legal system remains firmly rooted in the

principle of "human creation," and the refusal to

extend copyright protection to AI-generated works

has fueled the conversation between industry and

legal scholars. Many technology enterprises and

academics are now calling for the amendment of

existing laws to include AI generated content under

development and application of AI technologies

(Ropat, 2020).

3.2 The Legal Attitude and Practice in

the European Union

The European Union (EU) has treated the copyright

protection of AI-generated content in a cautious way,

and the core principles can be found throughout its

Council, 2019). According to the Directive, a work

must reflect "the author’s personality" and require

human intelligence and individualized expression

during the creation process. Judicial practice in the

EU has also emphasized this principle. For instance,

in the Eva-Maria Painer v. Standard Verlags GmbH

case, the court noted that originality should reflect the

creator’s freedom and individuality (Court of Justice

of the European Union, 2011). This case sets an

important reference for future disputes related to AI

generated content.

In practice, the EU emphasizes the importance of

human intervention. Generally, the generation of

AIGC is not considered as a copyright protected work

if it processes the information by the algorithms and

data and without any form of human involvement.

But if a user feeds the generative model creative

instructions, dynamically manipulates the outputs, or

edits the generated results, the content that results

may be copyrighted.

Moreover, the EU is looking at establishing new

regulatory requirements regarding attribution and use

of AI generated content. The 2020 White Paper on

Artificial Intelligence further clarified the standard

for human involvement in AI-generated content

(European Commission, 2020). The thinking is to

make sure this balances out protection of creators"

rights with enabling AI innovation. Through these

provisions and practices, the EU emphasizes the

central role of human creators in the creative process,

seeking to balance the protection of creators" rights

with the promotion of AI technology innovation.

3.3 The Legal Practice in China

In China, the current Copyright Law clearly requires

works to be original and created by natural persons,

legal persons or other organizations. This standard

makes it challenging in practice whether AI-

IESD 2025 - International Conference on Innovative Education and Social Development

generated content can be copyrighted. In recent years,

this issue has been widely discussed in Chinese

academic circles and judicial practice departments.

In the existing judicial practice, a case adjudicated

by Nanshan District People’s Court in Shenzhen at

the end of 2019 and widely reported in early 2020,

namely the case of Tencent Dreamwriter generating

financial articles, provides a reference for this issue

(Shenzhen Nanshan District People's Court, 2019). In

this case, Tencent created a financial news report by

using its AI tool "Dreamwriter" to automatically

generate the news report, and in this process, people

are used to bringing the topics, supplying data, editing

and publishing the news. The court found that

because of all this substantial human intervention and

creative input to AI-generated content, the final text

constitutes an “original” work subject to copyright

law.

This topic has also been examined in depth by

academic circles. In fact, some scholars argued it is

necessary to define and evaluate the creative labor

and decision-making input of humans in an AIGC

creation process for ensuring reasonable copyright

protection of AIGC works at legal level. At the same

time, the Chinese government has raised the strategic

importance of AI in policy manuals like the New

Generation of Artificial Intelligence Development

Plan (2017) and encouraged a better intellectual

property protection system (State Council, 2017).

Academic institutions, relevant research institutions

and the National Copyright Administration also

continue to organize discussions and exchanges to

provide references for future legislation and judicial

practice.

In general, judicial practice in China tends to

consider AI merely as a help in creation, rather than a

creator in itself for the creation of a work. This does

not mean that AI-generated content cannot be subject

to copyright protection as the AI-generated content

may in fact have substantial human contribution to

the creative process. This trend offers a first path of

future amendments or interpretations of related laws.

3.4 Attitudes in Other Countries

The United Kingdom: The United Kingdom is one of

the few countries in the world where AI-generated

content is treated very differently; under the

a developer’s copyright or a user operating an AI if

he or she has contributed "creatively" to the creation

(Leung, 2021).

Japan: Japan takes a relaxed attitude, encourages

the open use of AI-generated content, and takes a

more conservative stance on copyright issues, that is,

AI-generated content generally does not enjoy

protection can be achieved through other means

(Okuyama, 2020).

4 LEGAL INSIGHTS:

CONSTRUCTING A

FRAMEWORK ADAPTED TO

AIGC DEVELOPMENT

4.1 Redefining the Standard of

Originality

In view of the copyright disputes caused by AIGC, it

is particularly important to redefine the "originality"

standard in copyright law. To address the copyright

disputes arising from AIGC, redefining the standard

of "originality" in copyright law is particularly

important. Traditionally, originality requirements

focus on humans" intellectual labour and subjective

judgement in the creative process, however for the

content created by AI, its "creativeness" lies next to

the reorganization and creation of data by the

algorithm. In order to achieve that, copyright law may

expand its originality standard too and grant

In particular, the law may take "human

intervention" as one of the main criteria for judgment

in attribution to originality. The U.S. Copyright

Office’s guidelines for the registration of AI-

generated content are quite clear: human creative

contributions form a considerable part of obtaining

"Human intervention" is a key measure of originality

embodied in the official guidelines. Copyright

protection for such works should be possible as long

as it can be shown that humans guided, selected, or

supervised the creation of AI content. The re-

definition could push human AI co-creation and

create new avenues for more advancement in

technology.

4.2 Diversified Rules for Copyright

Ownership

The current controversy over the attribution of

Comparative Study and Insights on Copyright Protection of AIGC from an International Perspective

developers, users, or the public domain. In order to

better adapt to this new form of creation, the law can

try to construct diversified copyright ownership rules,

which are detailed according to different creative

scenarios.

AI Developer: If the AI-generated content relies

heavily on the developer's technical and algorithmic

design, the copyright can be vested in the developer.

Such rules can incentivize companies and

technologists to invest in the development and

optimization of AI systems.

Users: For users who have played an important

guiding role in the creative process, the law may

consider them as copyright owners. Such rules can

stimulate the enthusiasm of users to use AI for

personalized creation.

Public domain: If AI-generated content is

completely devoid of human participation and its

creation process can be replicated by anyone, then

such content should be put into the public domain for

free use by society. This will ensure that

technological achievements are shared and utilized to

the maximum extent possible by the public.

4.3 International Cooperation and

Coordination

As the application of AI technology is characterized

by globalization, countries should also coordinate the

avoid transnational legal disputes caused by unclear

First, under the umbrella of the WIPO (World

Intellectual Property Organization), there is a first

amendment of existing international copyright

treaties to make clear what AI generated content is

and what the appropriate standards are for protection

thereof. Second, countries can establish AIGC best

practice guidelines in copyright protection that all

countries can jointly use as a reference for national

legislation and justice so countries can implement

laws in accordance with technological innovation.

4.4 Legal Insights for China

By comparing the copyright practices of different

countries, China can improve the copyright protection

of AIGC in the following aspects:

Improve the legal system: In light of the actual

situation of technological development in China,

should be clarified as soon as possible, especially for

content created by AI and human collaboration, clear

identification standards and ownership rules should

be provided in the law.

Establish a rapid response mechanism: Due to the

wide application and rapid change of AI-generated

content, China’s copyright protection system has

established a corresponding rapid response

mechanism to respond to possible new copyright

disputes in a timely manner and ensure that the law

ADAPTS to the pace of technological development.

Promote the integration of technology and law: To

strengthen the legitimate use of AIGC content,

governments should foster cooperation between the

AI Industry and the legal community to establish

technical tools such as blockchain-based content

traceability systems to facilitate copyright protection

and monitoring (WIPO, 2020).

5 CONCLUSION

This paper discusses the copyright protection of

AIGC, especially from the perspective of originality

identification, copyright attribution and international

legal comparison. The emergence of AIGC poses a

major challenge to traditional copyright law, and the

current copyright protection framework faces

shortcomings in dealing with AI-generated con-tent,

especially because AI lacks human subjective

awareness and creative intent, which complicates the

determination of originality and authorhood.

Through international comparison, it can be seen

that different countries have adopted different

attitudes towards AIGC copyright protection, with the

United States and the European Union emphasizing

the direct creative input of human beings, while some

countries such as the United Kingdom and Japan

taking a relatively relaxed stance. Based on this

analysis, this paper puts forward some suggestions,

such as redefining the originality standard,

constructing diversified copyright ownership rules,

and strengthening international cooperation and

coordination, so as to promote the modernization of

REFERENCES

Abbott, R. 2018. Artificial Intelligence, Big Data, and

Intellectual Property: Protecting Computer-Generated

Works in the United States. IP Theory, 8(1).

IESD 2025 - International Conference on Innovative Education and Social Development

Cai, L., Yang, G. 2024. The Dilemma of Identifying AI-

Generated Content (AIGC) as Works and the Construc-

tion of Copyright Standards. Publishing and Distribu-

tion Research, (1): 67–74.

China Academy of Information and Communications Tech-

nology. 2022. White Paper on AI-Generated Content

(AIGC) [EB/OL]. China Academy of Information and

Communications Technology.

Court of Justice of the European Union. 2011. Eva-Maria

Painer v Standard VerlagsGmbH and Others (Case C-

145/10, ECLI:EU:C:2011:798).

European Commission. 2020. White Paper on Artificial In-

telligence: A European Approach to Excellence and

Trust. Brussels, 19.

European Parliament and Council. Directive (EU)

2019/790 of the European Parliament and of the Coun-

the Digital Single Market and Amending Directives

96/9/EC and 2001/29/EC. Official Journal of the Euro-

pean Union, L 130/92, 17 May 2019.

Policy: Trends and Perspectives. Journal of Intellectual

Property Law & Practice, 16(7): 721–731.

Li, B., Bai, Y., Zhan, X., Li, G. 2023. The Technical Char-

acteristics and Evolution of AI-Generated Content

(AIGC). Library and Information Knowledge, 40(1):

66–74.

Li, X., Hu, Y., Wang, M., et al. 2024. A Review of Artificial

Intelligence-Generated Content: Applications, Risks,

and Governance. Library and Information Work,

68(17): 136–149.

Intelligence: Current Legal Developments and Future

Prospects. Journal of Intellectual Property Law &

Practice.

Ropat, D. 2020. Authorship and Ownership of AI-Gener-

ated Works: The Need for Change in Copyright Law.

Journal of Intellectual Property Law & Practice,

15(11): 893–897.

Shenzhen Nanshan District People's Court. 2019. Civil

Judgment No. (2019) Yue 0305 Minchu 18180. China

Judgments Online.

State Council. 2017. The New Generation Artificial Intelli-

gence Development Plan [EB/OL]. [Cited on December

8, 2024]. Source: http://www.gov.cn/zhengce/con-

tent/2017-07/20/content_5211996.htm.

right Office Practices (3rd ed.). Sections 313.2,

503.03(b). Retrieved from https://www.copy-

right.gov/comp3/.

ance: Works Containing Material Generated by Artifi-

cial Intelligence.

quest for Reconsideration for Refusal to Register "A

Recent Entrance to Paradise" (Correspondence ID: 1-

3JLJA2G).

Kashtanova re: Zarya of the Dawn (Registration No.

TXu002368461). February 21.

Wang, C. 2024. On the Basis for Recognizing AI Users as

Shanghai University of International Business and Eco-

nomics, 31(06): 79–91.

World Intellectual Property Organization (WIPO). 2020.

Blockchain Technologies and IP Ecosystem – Whitepa-

per. Geneva: WIPO. Retrieved from

https://www.wipo.int.

Wu, H. 2024. On the Copyrightability of AI-Generated

Content: Practice, Theory, and Institution. China Law

Review, (03): 113–129.

Yu, C. 2024. The Rejection of Copyrightability for AI-Gen-

erated Works. Dispute Resolution, 10(9): 16–20.

Comparative Study and Insights on Copyright Protection of AIGC from an International Perspective