Exploring the Copyrightability of AI-Generated Content

Yanwei Huang

Chengyi College, Jimei University, Xiamen, China

Keywords: Copyright Law, Artificial Intelligence Generated Content, Copyrightability, Originality.

Abstract: “Artificial intelligence+” was included in the 2024 government work report of China for the first time,

marking the trend of scientific and technological innovation led by artificial intelligence, and demonstrating

China’s firm support for developing the industry chain of artificial intelligence. The works generated by

artificial intelligence not only present extraordinary artistic charm, but also have huge economic value. The

ideas indicated in the judgment of the first artificial intelligence case by Beijing Internet Court reveal that at

the moment when generative artificial intelligence is in the ascendant and becoming increasingly popular, it

is unreasonable to simply deny the originality of AI-generated works and their close connection to human

intellectual investment on the grounds that the creator is not a natural person. In the meantime, the exclusion

of artificial intelligence content from copyright protection may cause the copyright system exist in name only

in this era of rapid development.

1 INTRODUCTION

As a frontier field of scientific and technological

innovation, artificial intelligence (AI) has evolved

from the initial theoretical concept in science fictions

to the wide application in life since it was first

proposed in the 1950s. Developing at a remarkable

speed, AI has not only profoundly changed people’s

lifestyle, but also made a huge impact on the world’s

social and economic structure. Thanks to the

popularization of basic technologies such as high-

performance computers, high-speed networks and big

data, AI has become the focus of scientific research

and will continue to lead the technological revolution

and industrial upgrading in the future in order to meet

the urgent need of promoting the prosperity and

developing the global economy.

Generative artificial intelligence (or the text-to-

image models discussed in this article) represents a

major research progress in the field of AI in recent

years. This technology provides an easy and

convenient creative solution that can efficiently

convert the description in natural language by the

users into vivid images, breaking the limits of time-

consuming and labor-intensive traditional drawing

process and the constraints of personal painting skills

to bring in new possibilities for creators to realize

their ideas.

However, with its rapid development and

expanding application scope, generative artificial

intelligence has brought significant legal challenges:

should the content created by generative artificial

intelligence with human creativity (referred to as “AI-

generated content”) be considered as an object

protected by the Copyright Law? The works

stipulated in the Copyright Law of the People’s

Republic of China must meet the following four

requirements: “whether it is original; whether it is an

intellectual achievement; whether it is expressed in a

certain form; whether it falls under the realm of

literature, art, or science (Jing 0491 Min Chu No.

11279).”AI-generated content obviously meets the

latter two criteria, but there is still great controversy

in the academic community as to whether it is original

and whether it can be regarded as the result of human

intellectual labor.

On November 27, 2023, Beijing Internet Court

made a judgment on China’s first infringement case

of generative AI, which to a certain extent has

recognized the copyrightability of AI-generated

content, marking a breakthrough in resolving the

disputes over the ownership of AI-generated content

in China.

During the trial, the plaintiff accused the

defendant of removing the watermark from the

pictures he created and published on Little Red Book

using Stable Diffusion (a text-to-image model)

without authorization and then publishing it on

Baijiahao. The plaintiff claimed that during the

generation process, he obtained the satisfactory

Huang, Y.

Exploring the Copyrightability of AI-Generated Content.

DOI: 10.5220/0013975100004912

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 151-157

ISBN: 978-989-758-779-5

151

picture by selecting models, inputting prompt words,

and setting generation parameters, which reflected the

choice, selection, arrangement and design of the work

and demonstrated his intellectual labor. The picture

involved has been recognized by many users after

being released on the platform. Being recognized as

works by the standards of the general public indicates

that it has the originality. As a result, the picture

involved is qualified as the object protected by the

others to confuse the original author of the picture and

infringed the copyright. The defendant also claimed

to have no intention of infringement because the

picture involved was not used for commercial profit.

The defendant could not provide the specific source

of the picture or explain why the watermark was

erased; it was uncertain whether the plaintiff owned

the rights to the picture.

The court held that the plaintiff input the prompt

words in the Stable Diffusion model and requested

the generation of a close-up of a beautiful woman in

the light of dusk in a photographic style.

Subsequently, specific descriptions were added to the

initially generated pictures, and the parameters were

adjusted. Finally, a satisfactory picture was selected.

This process reflects the plaintiff’s intellectual

investment. During this process, the plaintiff obtained

the picture through continuous adjustment and

revision, demonstrating his aesthetic choice and

personal judgment. The generated pictures also have

identifiable differences from previous works, so the

picture is original. The court finally determined that

the defendant’s behavior constituted infringement

and ruled that the defendant should bear the

corresponding civil liability (Jing 0491 Min Chu No.

11279).

The judgment of this case not only provides

important reference and guidance for future legal

practices of this kind, but also raises the awareness of

all sectors to protect intellectual property rights

against the rapid development of cutting-edge

technology. In order to better adapt to the ever-

changing era of artificial intelligence, boost the

comprehensive prosperity of the market economy and

culture, and prevent the impact on the copyright order

of literature and art by the increasing AI-generated

content, it is necessary to discuss its characterization

in the Copyright Law. This article will discuss the

view that denies the copyrightability of AI-generated

content through arguing that AI-generated content is

the result of human intellectual investment and has

originality by combining the judgment of the first AI

case from Beijing Internet Court, aiming to contribute

to the construction of a more complete and reasonable

2 THE DENIAL OF THE

COPYRIGHTABILITY OF

AI-GENERATED CONTENT

AND ITS PROBLEMS

2.1 The Denial of the Copyrightability

of AI-Generated Content

Whether AI-generated content can be regarded as

works is challenging and impacting the framework of

traditional Copyright law, triggering thoughts and

discussion in the academic community. Scholars hold

different views on this, which are mainly affirmative

views (recognizing the copyrightability of AI-

generated content) and the opposite.

The view denying the copyrightability of AI-

generated content claims that it is not human

intellectual achievements and lack originality, and its

constituent elements do not meet the definition of

works in the Copyright Law. Although similar to

human works in terms of expression, AI-generated

content is not essentially objects protected by the

identity of the author and the ownership of copyright.

For example, some scholars have cited the classic

case of “monkey selfie dispute” in their articles,

emphasizing that the objects protected by the

achievements.

The aforementioned view believes that although

the photos taken unintentionally by macaques are

difficult to distinguish from professional photographs

in quality, given that macaques are not natural

persons, their eye-catching selfies are not protected

by the Copyright law. There can be proviso in the law,

but there is no exception in logic. The nature of

generative AI is no different from the example of

“monkey selfie”, which is not the intellectual

investment of natural persons (Wang, 2023). The

current generative AI content is mainly based on

algorithms, rules and models. Its creation process is

essentially the result of executing preset procedures

and methods instead of independent and personalized

creation of the author. The process of generating

content does not involve the intelligence required for

creation (Wang, 2017). In summary, AI-generated

content is not protected by the Copyright law.

The author disagrees. The view denying the

copyrightability of AI-generated content tends to

exclude it from the scope of copyright protection,

IESD 2025 - International Conference on Innovative Education and Social Development

152

which fails to guide the development of advanced

productivity, and also makes the copyright system

face severe challenges in the upcoming AI era. In this

regard, this article will explain the necessity for

relevant laws to include AI-generated content as

works and to entitle them to copyrightability.

2.2 AI-Generated Content as Human

Intellectual Achievements

A photographer walked into the jungle of Indonesia

with a professional camera. By chance, a group of

curious macaques discovered the camera on the tripod

and fiddled with it. The macaques inadvertently

pressed the shutter hundreds of times. It was this

series of accidental operations that captured the

brilliant smile of a celebes crested macaque,

presenting a photo that looks like a human selfie.

The macaque may have never expected that this

unconscious behavior driven by animal instinct

would become an indispensable topic in the study of

the Copyright law. Since 2015, People for the Ethical

Treatment of Animals (PETA) has sued the

photographer on behalf of macaques, claiming that

the macaque should fully enjoy the copyright of the

selfie. They accused the photographer of

infringement by using the photo without permission

and making continuous profits. After two years of

litigation, the judge believed that the relevant laws

protecting copyrights do not apply to animals, and

finally ruled that the photographer enjoys the

What if the photographer of that picture wasn’t a

monkey? Imagine if the situation was even more

bizarre at the time: a cricket jumped onto the shutter

of the camera and accidentally “created” a

masterpiece with excellent composition and color. I

guess no one would step up and claim that the cricket

created the photo. The ownership of the copyright of

the photos taken by macaques has become a focus of

controversy, but there is no need to explore whether

the cricket can be a photographer. What is the

essential difference between the two examples?

The author believes that there is no difference

between the two. Whether it is a monkey or a cricket,

they are both animals, so the photos they “took” are

completely products of chance. The philosopher Karl

Marx pointed out that the most essential difference

between humans and animals is that humans have

consciousness and can actively create and change the

world, while animals have no consciousness and

cannot exert subjective initiative. The macaques

fiddling with the camera and the cricket jumping are

just simple conditioned reflexes instead of their input

for the purpose of “intellectual achievements”. This

is like making a monkey sit in front of a typewriter.

Even if it successfully typed out a copy of Hamlet,

neither the monkey nor the individual who arranged

this would be considered to have plagiarized

Shakespeare's works.

However, AI is not an animal. The infinite

potential and powerful strength of the code are the

imitation and reproduction of human mind. In terms

of the internal operating principle, AI not only

reproduces the human thinking system, but also has

the ability to imitate and even replace human

intellectual investment. Computer hardware, as the

basis for AI operation, corresponds to the lowest level

of physiological processes in human intellectual

investment; AI programming languages reflect the

primary information processing links in human

intellectual investment; AI programs involves the

highest level of human intellectual investment –

thinking strategies (Yang, 2021). Taking Stable

Diffusion as an example, the principle of image

generation by this deep learning model is to achieve

the conversion from text description to detailed image

through diffusion model. The text encoder accurately

analyzes the text information input by the user and

converts it into internal instructions; the neural

network noise predictor based on the UNet

architecture is the core of Stable Diffusion. In the

process of image generation, it first receives a noisy

picture, then gradually removes the noise to produce

a clear, creative and brand-new pattern that conforms

to the user’s description; the variational autoencoder

(VAE) realizes image compression through Encoder

and Decoder (Xu, 2024). Users can make the

generated pattern more in line with their own

creativity by introducing techniques such as

conditional control and feature embedding.

Therefore, it undoubtedly contradicts the logic of

humanism to compare AI creations that reproduce the

complexity and creativity of human brain activities

with the conditioned reflex behavior of macaques

photographing under external stimuli.

In addition, some scholars argue that the

photographer locate the camera on a tripod at a

specific spot in the jungle, which reflects human

intellectual achievement. This argument lacks

theoretical basis, since there is no connection between

where the camera is put by the photographer and the

generation of the monkey selfie. The macaque

snatching the camera and pressing the shutter is pure

accident. The human intellectual investment of

arranging the camera did not play an active role in the

generation of the photo. In this case, neither the

intellectual achievements of human investment were

Exploring the Copyrightability of AI-Generated Content

153

shown, nor was there the possibility of considering

the macaque as the photographer.

In contrast, users only need to run Stable

Diffusion on their computers and input specific text

information into the text encoder while adding control

conditions to personalize the generated content in

order to create an image that reflects their inspiration.

The more detailed the input text information is, the

more accurately the generated image can reflect the

user’s creativity. This clearly reveals that there is a

fundamental difference between AI creation and

macaque photography. The process of AI creation

relies on the investment of human intelligence. As

Beijing Internet Court pointed out in its judgment:

“From the time the plaintiff had an idea about the

picture to his final selection of the picture involved,

the plaintiff did some intellectual investment, such as

designing the presentation of the character, selecting

prompt words, arranging the order of prompt words,

setting parameters, and selecting the picture that he

wanted (Jing 0491 Min Chu No. 11279).” In this case,

the carefully designed experiments by the plaintiff

fully demonstrated the close connection between

human intellectual investment and the quality of

creation, proving that it was the picture in this case is

the result of the plaintiff’s intensive intellectual labor.

The judgment opposes the view that AI creation is

comparable to macaque photography, which is not

only an affirmation and respect for the concept of

Humanism and the hard work of the creators but also

consistent with the legislative purpose of the

2.3 The Manifestation of Originality of

AI-Generated Content

In terms of originality, there is a view that cannot be

ignored: since generative AI technology is written

based on algorithms, and the creation process follows

the steps of fixed models, it lacks the unique

character, emotional appeal and even spirit that

human creator can put into a piece of work.

Therefore, the process of generating content by AI

does not belong to the category of intellectual

creation and cannot meet the requirement of

originality. In addition, some people even refer to the

process of AI creation based on massive material

training and regeneration as “corpse collage”, which

is quite Ludditism-like when manual workers

destroyed the textile machines during the Industrial

Revolution. In this regard, it is important to discuss

what is the emotional appeal of the work. What are

the irreplaceable qualities in human creation that

distinguish it from the works generated by artificial

intelligence?

When discussing aesthetic theory, transference is

a key concept, which reveals the intrinsic connection

between human emotions and aesthetic experience.

Transference refers to the individual projecting their

own emotions, thoughts or will onto external objects,

thereby producing a special experience as if the

external objects also have the corresponding

emotions or characteristics.

The poetry of Tang Dynasty and the lyrics of Song

Dynasty are the artistic treasures of Chinese culture,

in which the technique of expressing one’s aspirations

through objects is common. For example, on his big

day of becoming the government official, Meng Jiao

wrote with great joy: “Successful, faster runs my

horse in vernal breeze, I’ve seen within one day all

flowers on the trees (Xu, 2021).”as if the spring is

prospering together with him. Feng Yen-ssu used a

straightforward tone to depict a different spring

scenery: “Beautiful, this view of hills and streams, As

always along the roads of Chin-ling town; But youth

turns to age all the same (Translated by Daniel

Bryant).” which makes people feel sad about how

time flies. When facing the sunset, Chinese people

will mourn for the elapsed time; when seeing the

moon, Chinese people will miss the hometown in the

beam shed by this same moon. Zhou Dunyi loves

lotus because of its elegant temperament. Li He wrote

a poem to praise horses because the old horse is still

running thousands of miles just like a hero with

unfulfilled ambitions in his later years.

Are these complex emotions really related to the

nature of the things and images mentioned above? If

so, anyone should feel the same when looking at the

same rose, and the emotional barriers between

humans will no longer exist. When one gazes at the

rose, it may has a yearning for love within; but others,

seeing the crumbling petals, may contemplate the

impermanence of life. Humans’ emotion originates

from one’s own subjective feelings. Through the role

of transference, humans attach emotion to what they

see. This is the aesthetic empathy, one of the most

influential schools of modern Western aesthetics,

argued by the German psychologist Lipps: the

aesthetic object will stir up certain experience in the

individual’s mind, and then project these internal

experiences onto the external object, thus forming a

kind of resonance.

In other words, when people appreciate a thing of

beauty, whether it is the wonders of nature or

fascinating literary works, the emotions experienced

are not the inherent characteristics of the object being

appreciated. Instead,when the memories and

IESD 2025 - International Conference on Innovative Education and Social Development

154

experiences rooted in the subconscious are touched,

through a series of specific complex psychological

activities and thinking processes, the specific

emotions aroused by themselves are transferred to the

aesthetic object.resulting in the misunderstanding that

the object contains such emotions.In other words,

emotions are produced by people, not by what people

see..People are appreciating themselves (Wei, 1999)!

When Zhuangzi standing on the bridge, what he

appreciated was not the happy fish, but the open-

mindedness and freedom in his heart.

If the “emotional appeal endowed in the work” is

regarded as the key to intellectual creation, it does not

depend on the creator’s investment, but on the

audience’s experience. Otherwise, all the audience

around the world should have the same emotional

resonance with what the creator has invested when

reading the same article. This denies the diversity of

human emotions. On the video platform called

Bilibili, there is a video called “Mohe Dance Hall

story painted entirely by AI made me cry!!!" (video

number: BV1Sd4y1c78G). The creator used

generative AI to carefully draw the illustration for

each line of the lyrics for the song Mohe Dance Hall,

vividly depicting the artistic images of the song. The

audience have been touched by the harmonious

integration of music and visual art. They think it is

hard to believe that these vivid images are all created

by AI. Although AI does not have consciousness and

cannot understand the story from a human

perspective, the work produced by its algorithms

according to a fixed model managed to provide the

audience with artistic enjoyment. When the subject’s

soul is touched by feeling provoked by the object, and

the strong emotions is projected on the object, it can

be said that these creations contain rich emotional

appeals, which demonstrates that we can be moved by

AI works.

In traditional literary and artistic creation, the

creator’s unremitting pursuit and dedication are

crucial. However, is there a standard to measure the

depth of emotion that the creator has poured into the

artwork? Of course not. Even human artist may create

a masterpiece with just a few strokes at a flashing

moment of inspiration. However, sometimes the final

result is worthless even with great efforts. It is

obviously not objective to use “the creator’s

emotions” as a standard to judge intellectual creation.

Whether the work has emotional appeals should be

determined by the audience! We should not assume

from an unquantifiable perspective that AI works do

not have the possibility of containing profound and

rich emotional appeals.

In fact, whether or not to recognize the originality

of AI creations does not affect the application of

generative AI to all the fields of literature and art in

the digital age, including film, music, and literature.

AI is everywhere. For example, the animation “The

Dog & The Boy” produced by Netflix is a project of

using artificial intelligence to create all the animation

scenes, demonstrating a unique charm in both visual

performance and creative conception that has been

well-received by the audience. Another example is an

advertisement released by Ito En Co., Ltd., which

presents a female model generated by AI. In the film,

an elderly lady is holding the green tea and jumps

joyfully to raise the bottle to the camera, saying the

line with a dynamic tone: “My future starts from

now.” As soon as she finishes the line, she moves the

bottle away, and suddenly returns to youth. The

wonderful performance of the AI model vividly

presents the healthy, natural and refreshing selling

points of the green tea beverage. The packaging

design of this green tea product is also tailored by

generative AI. In 2023, Japanese cartoonist

Sandrovich Yabako frankly revealed on the social

platform that generative AI was used in the creation

of his work Kengan Ashura. It is unexpected that an

illustration that has sparked heated discussions

among netizens and was mistakenly believed to be

created by AI is actually due to a mistake in manual

drawing, while the content that was truly created with

AI has been ignored.

The Copyright Law of People’s Republic of China

has relatively loose requirements for the originality of

works. The requirements to reach a certain aesthetic

level or have specific academic value is not strict. As

long as the work is independently completed and not

plagiarized from others’, it can meet the standards and

naturally be protected by law. Historical materialism

emphasizes that the masses are the creators of history,

and the law should not go beyond the common sense

of the people. With the wide application of generative

AI in the fields of literature and art, the public is not

let down by AI-generated content because they lack

emotional appeals, or recognize any similarity

between AI-generated content and existing works.

Furthermore, the commercial advertisements created

with the generative AI has also managed to meet the

standard of “independent completion” and none of

them have been questioned in terms of originality or

caused of any infringement disputes such as

plagiarism. Then why should we doubt the originality

of AI-generated content? Will the content lacking

originality be attached with great value by the

practitioners in the field of entertainment, and

become the popular creative material to be well-

Exploring the Copyrightability of AI-Generated Content

155

received by the public? The answer is self-evident.

The view that AI-generated content is not original

contradicts to the practice in real life.

3 CONCLUSION

At present, although generative AI is not yet mature,

its influence continues to expand, and it is constantly

iterating and updating at an astonishing speed. The

types, styles and quality of generated content are

becoming closer to perfection. It is not even a fantasy

to replace humans in literary and artistic creation in

the future. If the Industrial Revolution liberated

humans from heavy physical labor, AI, with its

immeasurable potential, may liberate humans from

difficult mental activities (Jiang and Xue 2022). As

Beijing Internet Court stated in the judgment: “The

development of technologies and tools require less

human investment, but the copyright system should

remain in use in order to encourage the creation of

works (Jing 0491 Min Chu No. 11279).”

The Copyright law protects works and encourages

creators to obtain rich economic benefits, which in

turn stimulates them to create more high-quality

works to form a virtuous circle and promote the

overall prosperity of the cultural market. If we insist

that AI-generated content cannot be regarded as

works, but merely content generated by specific

algorithms in accordance with mechanical steps, it

means that the content, like the “multiplication table”,

belongs to the public domain and is not protected by

the copyright law. The copyright system will also face

the risk of being eliminated as AI becomes the

dominant force in literary and artistic production in

the future (Yi, 2017).Therefore, we need to consider:

can we bear to live in a world without the protection

of the copyright law?

In such an environment, the hard work and

wisdom of creators cannot be respected and protected

as they should be. The market will be flooded with a

large number of low-quality and uncreative copies,

while the truly valuable content will find it difficult

to stand out. If this continues, the cultural ecology of

the entire society will suffer a heavy blow, leading to

declining innovation ability and decreasing cultural

richness, and eventually the impoverishment of

civilization.

The torrent of history rushes forward, and the leap

of productivity is unstoppable. Looking back at the

development of human society, every major

breakthrough in science and technology is like the

beacon to illuminate the way forward for civilization

and leading it to a more glorious future. From the

roaring steam engines to the bright electric lights,

from the birth of the Internet to the application of AI

in the daily life of thousands of households, every

great invention is the crystallization of collective

human wisdom. Nevertheless, science and

technology cannot develop without a hitch, and the

changes it brings are often accompanied by

challenges to ethics, legal systems and even social

structures. In response to the problems caused by

artificial intelligence and the massive content it

generates, the copyright system should be sufficiently

forward-looking and adaptable, keep pace with the

times with an open and inclusive attitude, and

improve itself according to the internal regulatory

mechanism, rather than blindly excluding emerging

innovations.

In fact, artificial intelligence is an extension of the

human body, an externalization of human practice,

which carries the infinite possibility of liberating

humans from their lifestyles and production methods.

As long as the audience of AI creation is human,

human subjectivity will not be eliminated. No matter

how time changes, the brilliance of human

subjectivity will forever shine in the galaxy of history.

The copyright law will continue to protect the field

that humans have worked hard to cultivate day and

night.

REFERENCES

Jiang, L.D. & Xue, L. 2022. Era Challenges and Paradigm

Transformations in the Governance of China's New

Generation of Artificial Intelligence. Journal of Public

Management 19(02): 4.

The case of Li (the plaintiff) v. Liu (the defendant) over in-

fringement of the right of authorship and the right of

dissemination on the information network, Beijing In-

ternet Court Civil Judgment (2023) Jing 0491 Min Chu

No. 11279.

Wei, X.L. 1999. On the Origin of the Empathy and Its Role

in Literary and Artistic Creation. Journal of Beijing

Normal University 3: 74.

Wang, Q. 2017. On the Characterization of Content Gener-

ated by Artificial Intelligence in Copyright Law. Sci-

ence of Law 5: 150-152.

Wang, Q. 2023. The Qualitative Analysis of Content Gen-

erated by Artificial Intelligence in Copyright Law.

Tribune of Political Science and Law 41(4): 69-70.

Xu, S.H. 2024. Exploration and Utilization of AIGC in New

Media Content Production. Radio

＆

TV Broadcast En-

gineering 51(11): 13-14.

Xu, Y.C. 2021. Reflections and Recollections. In X.L. Liu

& X. Zhang & Q.Y. Wang & W.J. Zhang & X.T. Zhao.

& J.F. Bi (eds), Three Hundred Tang Poems Translated

IESD 2025 - International Conference on Innovative Education and Social Development

156

by XU Yuanchong 2: 664-665. Bejing: China Transla-

tion & Publishing Corporation.

telligence Products. Modern Law Science 43(4):104.

Yi, J. M. 2017. Are Artificial Intelligence Creations Works?

Science of Law 5: 144.

Zhu, G. 2024. Research on Legal Attributes and Ownership

Rights of AI-Generated Images. Intellectual Property

01: 27.

Exploring the Copyrightability of AI-Generated Content

157