Finding of Fact of Copyright Infringement in Artificial Intelligence

Generated Objects

Qing Luo

Law School, Capital University of Economic and Business, Beijing, 100071, China

Keywords: Intellectual Property Right, Artificial Intelligence, AI Generated Objects, Copyright Infringement.

Abstract: With the development of generative AI technology, different forms of AI generated objects appear in various

fields of people's lives. In the past, copyright protects the intellectual achievements and innovative research

of creators, but nowadays, it is difficult to identify and trace the specific behavior of the copyright subject and

infringement of AI, which leads to a series of AI fair use problems and copyright infringement problems. This

paper analyzes the rules and dilemmas of copyright infringement identification from the characterization of

AI generators, and proposes a mechanism for preventing copyright infringement by AI, which is expected to

promote the soundness of the legal regulation of AI in the whole process of training, reasoning and generating.

1 INTRODUCTION

Artificial intelligence is one of the most influential

technologies today, and it is changing people's way of

life and way of work in a very deep way. In the field

of artistic creation, in the field of scientific research,

as well as in the field of news writing and film and

television production, the variety and quantity of AI-

generated works are increasing, and the proportion of

AI-generated works in these fields, such as the

cultural industry, is also increasing. Against this

background, conflicts and contradictions between AI-

generated works and traditional copyrights are slowly

emerging, and because of the existence of such

conflicts, disputes over copyright infringement have

become particularly numerous. Accurately

determining copyright infringement of AI-generated

objects is related to whether the legitimate rights and

interests of creators, users, and related industries can

be protected and also affects whether the AI industry

and the cultural industry can develop in a healthy and

sustainable manner. This is of crucial significance in

maintaining a favorable intellectual property

ecosystem. The purpose of this paper is to build a

clarifying the considerations for the determination of

analyzing the difficulties in applying the rules of

infringement.

2 DIFFICULTIES IN

DETERMINING COPYRIGHT

INFRINGEMENT OF

ARTIFICIAL INTELLIGENCE

GENERATED OBJECTS

2.1 Fuzzy Definition of

Copyrightability of Artificial

Intelligence Generated Objects

In the traditional copyright law, the subject of

creation of works generally refers to natural persons

and requires works to have a certain degree of

originality to reflect the input of human intellectual

creativity; however, the process of generating

artificial intelligence and traditional human creation

is completely different. Artificial intelligence systems

are able to generate content with a certain degree of

expressiveness and value, such as paintings, music,

and articles, without direct human creative

intervention, by virtue of training on a large amount

of data and complex algorithms, which has led to

disputes about whether AI-generated content is

copyrightable (Yang, 2024). If the copyrightability of

AI-generated content cannot be clarified in the law,

then there will be a lack of basic prerequisites for the

determination of copyright infringement, and it will

be difficult to determine whether it can be protected

by the copyright law, and it will be difficult to

Luo, Q.

Finding of Fact of Copyright Infringement in Artiﬁcial Intelligence Generated Objects.

DOI: 10.5220/0014381600004859

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

In Proceedings of the 1st International Conference on Politics, Law, and Social Science (ICPLSS 2025), pages 387-390

ISBN: 978-989-758-785-6

387

determine whether it is possible to infringe on the

rights and interests of others (Hua, 2024).

2.2 Difficulty in Applying the

Determination Standard of

Infringement Behaviour

In the traditional copyright infringement

determination process, the general use of the "contact

plus substantial similarity" determination standard,

specifically, if the defendant has access to the

plaintiff's work and after comparison, it can be found

that there are substantial similarities between the two,

then it is likely to be an infringement. However, the

entire process of creation of AI-generated works has

a very high degree of complexity and non-

explainability, and the correlation between the

content it generates and its training data is not the kind

of simple and direct causality, so it is necessary to

judge whether the AI-generated works have been

exposed to other protected works intuitively, just as

in the case of traditional works, and whether the

similarities are due to independent creation or

coincidence, or whether the similarities are due to

independent creation. Creation or substantial

similarity due to infringement, it is very difficult to

do so (Lin, Zhang, Shi, et al, 2025). In addition, AI is

likely to incorporate multiple different data sources

and creative styles, which makes it more difficult to

determine whether it constitutes an infringement or

not (Yang, 2024).

2.3 Controversy over the Definition of

the Subject of Liability

When the content generated by AI constitutes

complicated and critical thing to define the specific

responsible body. The developer, user, and owner of

the AI system may be related to the infringement

responsibility. The developer's main job is to be

responsible for the design of algorithms and models

and to provide the necessary infrastructure for the

content generated by the AI. Developers are

responsible for designing algorithms and models and

providing the infrastructure required to generate

content for the AI; users input commands and data

into the AI as needed to guide the AI in its creative

activities; and owners may have the so-called right to

control and benefit from the AI system and the

content it generates (Su, 2024). In different

application scenarios and legal relations, these

different subjects in the infringement of the role

played by the degree of fault are difficult to accurately

and accurately divide, so that when the need to pursue

the responsibility of infringement arises, there is no

way to determine who should bear the responsibility

and the responsibility can not be reasonably and fairly

distributed, which has a bad influence on the accuracy

and fairness of the determination of copyright

infringement. This has a bad influence on the

accuracy and fairness of the determination of

3 CONSIDERATIONS FOR THE

DETERMINATION OF

OF ARTIFICIAL

INTELLIGENCE GENERATED

OBJECTS

3.1 Application of the Principle of

Technological Neutrality

Technology itself is neutral, the purpose of the

development of artificial intelligence technology is to

improve the efficiency of work, create new value, it

itself is no legal sense of good or evil attributes, in

determining whether there is a copyright

infringement of the content generated by artificial

intelligence, the principle of technological neutrality

should be taken into account, to prevent unnecessary

inhibition of the use of technology on the

development of innovative activities and industries.

Inhibitory effect on the development of innovation

and industry due to the use of technology (Zhan,

2024). If the developer or user of the AI system has

used the technology without subjective fault and in

accordance with reasonable technical specifications

and common industry practices, it may be necessary

to look elsewhere for the cause of the infringement

and the responsible party, rather than simply

assigning responsibility to the technology itself or the

person using the technology. or the person who uses

the technology (Mei, 2024).

3.2 Judgement of the Degree of Fault

In the process of determining copyright infringement,

fault is a very key consideration, when it comes to the

infringement of AI generators, we need to

comprehensively consider the degree of fault of each

relevant subject, for example, whether the developer

has taken the necessary measures to avoid infringing

the copyright of others when designing and training

ICPLSS 2025 - International Conference on Politics, Law, and Social Science

388

the AI system, which includes the need to carry out a

rigorous examination of the legitimacy of the training

data, and also the need to use the training data to avoid

infringing the copyright of others. For example, when

designing and training the AI system, whether the

developer has taken necessary measures to avoid

infringing others' copyrights, which includes strict

examination of the legitimacy of the training data and

adopting reasonable data processing methods;

furthermore, whether the user has used the content

generated by the AI in the case that he/she knew or

should have known that his/her behaviour might

constitute infringement, and whether there has been

inappropriate intervention in or misuse of the AI

system, and so on. According to the different degrees

of fault, the infringement responsibility that each

subject should bear should be reasonably determined,

so as to do so can make the infringement

determination to achieve fairness and reasonableness

3.3 Clarification of the Boundary of

Fair Use

The fair use system in the copyright law mainly

balances the interests of the creator and the public and

the establishment of the relationship. This system

allows, in some specific circumstances, a certain

degree of use of the protected works, and does not

need to specifically go to obtain authorization

(Murray, 2023). In the case of the determination of

intelligence generation, we need to figure out where

the boundaries of fair use in the end. It is clear that

the use of other people's works by AI systems belongs

to the scope of fair use or not. For example, if an AI

system analyses and processes a relatively small

number of works for the purpose of study and

research, it may be considered fair use; however, if it

copies and uses other people's works on a large scale

without authorisation to train the AI model, and then

generates commercial works with a high degree of

similarity to the original works, it may have exceeded

the scope of fair use. However, if the work is copied

and used on a large scale without authorisation to

train the AI model, and then generate a commercial

work with a high degree of similarity to the original

work, then it may have exceeded the scope of fair use

and constituted copyright infringement (Yao, 2024).

4 THE CONSTRUCTION OF

PREVENTION MECHANISM

OF ARTIFICIAL

INTELLIGENCE

GENERATORS

4.1 The Application of Technical

Means

With the help of technical means to prevent copyright

infringement of AI generated products, this is a very

key initiative, to develop and use effective digital

watermarking technology, for example, can be the

author's name, copyright information into the content

generated by the AI, so that in the subsequent

dissemination and use of the process, can accurately

identify where the work is from and who owns the

encryption technology can be used to protect AI-

generated content with high commercial value, so as

to prevent unauthorised copying and distribution, and

a traceability system can be created for AI-generated

content. This system will record the key information

in the generation process, such as where the data used

in training came from, what are the specific

parameters of the algorithm, etc., which can provide

very strong technical support for the determination of

infringement, and can also protect in terms of

evidence (Su, 2024).

4.2 Improvement of Legal Norms

The legislative and judicial authorities need to

improve the relevant legal norms in a timely manner,

the purpose of doing so is to adapt to the new needs

arising from the protection of copyright in the era of

artificial intelligence, to clearly define the

copyrightability standards of AI-generated materials

and the rules of copyright attribution, with the help of

which can provide a clear and explicit legal basis for

the determination of infringement, in addition to one

point, it is also necessary to refine the criteria for the

determination of copyright infringement. In addition

to one point, it is also necessary to refine the

according to the characteristics of artificial

intelligence to formulate special rules and a detailed

operation process, and to reasonably define different

subjects in the artificial intelligence copyright

infringement of the responsibility of the way, and the

specific proportion (Zhu, Cui, Wang, et al., 2024). By

Finding of Fact of Copyright Infringement in Artiﬁcial Intelligence Generated Objects

389

virtue of such a way, to ensure that the pursuit of

infringement responsibility has fairness and

operability, in addition, it is also necessary to

strengthen the exchange and cooperation between

countries in the legal aspects, and actively promote

the formation of a unified legal framework for the

protection of artificial intelligence copyright, so as to

be able to face the complexity and variability of

transnational application of artificial intelligence

technology (Chalu, 2024).

4.3 Ethics and Industry

Self-Regulation Guidance

In addition to relying on legal and technical means,

moral norms and industry self-discipline also play an

indispensable role, people should actively advocate

those enterprises and institutions related to artificial

intelligence and practitioners to establish a correct

awareness of intellectual property rights, fully respect

all the creative achievements of other individuals, and

consciously comply with copyright laws and

regulations. People need to formulate self-regulatory

guidelines and codes of conduct for the AI industry,

and strongly encourage enterprises to take various

positive and effective measures to prevent the risk of

use of AI technologies, such as establishing a set of

their own internal copyright review mechanism and

conducting copyright-related training. People also

need to strengthen the public's concern and

understanding of the copyright of artificial

intelligence, with the help of this way to improve the

public's awareness of copyright protection, and then

create a good atmosphere of social opinion, so as to

constrain the emergence of copyright infringement of

artificial intelligence generators from the moral level.

5 CONCLUSION

As a focal point of copyright system change in the

digital era, the infringement determination of AI-

generated works is essentially a dilemma between

technological innovation and the existing legal

framework. Through the analysis of legal

hermeneutics and comparative law, this study finds

that the core contradiction in the determination of

infringement focuses on the three dimensions of the

judgment of the eligibility of the creative subject, the

blurring of the originality standard, and the breakage

of the chain of responsibility. The specific path

should follow the principle of “three-step

progression”: introduce the “limited copyright”

system at the stage of rights allocation, and include

AI generated objects into the scope of neighboring

rights protection; build the composite standard of

“substantial similarity + contact possibility +

algorithmic independence” at the stage of

infringement determination; and implement the

attribution model of “presumption of fault +

proportionality” in the mechanism of responsibility

sharing. This is not only a systematic response to

build a digital copyright governance system, but also

a key breakthrough point to improve the rule of law

governance in the field of emerging technologies,

which has important theoretical and practical value

for promoting the high-quality development of the

digital economy.

REFERENCES

Hua, J. 2024. Brain-computer interface generates content

Social Science (09):115-124+147.

Huang, L. 2024. Behavioral Regulation Model of Artificial

Intelligence Generated Content Copyright Protection--

Taking Sora Wensheng Video as an Example.

Journalism (06):59-73.

Lin, J. H., Zhang, Y. F., Shi, P. C., et al. 2025. The rise of

algorithmic marketplaces: concepts, challenges and

future developments. Southern Economy (01):1-17.

Infringement and Fair Use. SMU Science and

Technology Law Review 26(04):259-354.

Su, Z. Q. 2024. On the Paradigm Shift in the Recognition

of Copyright Objects. Intellectual Property Rights

(02):67-86.

Wu, H. D. 2024. On the Copyrightability of Artificial

Intelligence Generated Content: Practice, Jurisprudence

and Institution. China law review (03):113-129.

Yang, M. H. 2024. Artificial Intelligence Legal

Relationship Challenges and Institutional Adjustment--

Taking the Wave of U.S. Local Legislation as an

Example. Journal of China University of Political

Science and Law (05):244-256.

Yao, Z. W. 2024. Determination of Artificial Intelligence

Generator Copyright Infringement and Its Prevention-

Centered on the World's First Infringement Judgment

of Generative AI Services. Local Legislation Research

9(03):1-17.

Zhan, L. 2024. Re-exploration of the status of legal mimetic

subject of artificial intelligence - oriented to the legal

subject system under the perspective of mimetic

philosophy. Journal of Beijing Institute of Technology

(Social Science Edition) 26(04):152-163.

Zhu, G., Cui, G. B., Wang, Q., et al. 2024. Is Artificial

Intelligence Generated Content (AIGC) Protected by

ICPLSS 2025 - International Conference on Politics, Law, and Social Science

390