Copyrightability of Artificial Intelligence Prompts

Miao Wu

Jurisprudence (Civil and Commercial Law Oriented), Law School, East China University of Political Science and Law,

Shanghai, 201620, China

Keywords: Natural Language Processing, Artificial Intelligence, Prompt, Copyrightability, Copyright Law.

Abstract: The rapid development of artificial intelligence (AI) models in natural language processing has transformed

prompts from purely technical tools into commodified digital assets with substantial commercial value. This

transformation has given rise to a burgeoning prompt trading market and triggered debates surrounding the

copyrightability of prompts. This paper examines whether and how prompts may be eligible for copyright

protection, emphasizing the necessity of constructing an institutional framework to safeguard the legitimate

rights of prompt creators and users, support the trading market, and facilitate the sustainable advancement of

AI technologies. To that end, the paper undertakes a comprehensive analysis of the essence, features, and

internal structure of prompts, aiming to assess the rationality and legal feasibility of affording copyright

protection under existing legal doctrines. Specifically, it focuses on three core aspects: the textual expression

constituting originality, the qualification and role of the author, and the intellectual labour embedded in the

creation of prompts. Moreover, this paper proposes enhancing the legal response by extending judicial

interpretation to establish clearer standards. It further advocates for a novel paradigm of human-computer

collaboration and the strengthening of industry self-regulation, thereby striking a balance between

incentivizing innovation and preventing the monopolization or abuse of digital rights.

1 INTRODUCTION

With the development of artificial intelligence (AI) in

the field of natural language processing, the threshold

of generative artificial intelligence has been gradually

reduced, which also makes the significance of AI

prompts in human-computer interaction continue to

rise. It is more well-structured and professional,

morphing from a simple technical command tool into

a text collection with originality, professionalism, and

functionality, giving rise to the emergence of the

prompts trading market. PromptBase, a platform that

provides artificial intelligence prompts trading, and

customization services, for example, attracts more

than 300,000 users according to its official website,

indicating the great market potential and massive

economic value of prompt engineering and prompt

creation. However, despite the fact that the prompt

trading market has begun to take shape on an

international scale, the legal positioning and rights

protection of prompts have not yet been clarified in the

legislation and judicial practice of various countries.

Corresponding author

In the increasingly prosperous situation of prompt

trading, it is urgent to give prompts a clear position

and protection, to provide a guarantee for the

development of prompt authors and the trading

market, and also provide a solid foundation for the

development of artificial intelligence technology. In

this regard, the article analyzes the copyrightability of

prompts by combining the nature and characteristics

of prompts, and proposes a path of identification,

which further provides ideas for the optimization of

international prompt copyright protection.

2 ANALYSIS OF THE

JUSTIFIABILITY AND

CHALLENGES OF

FOR PROMPTS

In practice, prompt transactions have formed a market

of a certain scale. This part explores the basis of its

Wu, M.

Copyrightability of Artiﬁcial Intelligence Prompts.

DOI: 10.5220/0014389800004859

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 559-564

ISBN: 978-989-758-785-6

559

possible copyright protection through the sorting out

of its nature and characteristics, as well as the analysis

of the legislative purpose of the copyright law and the

logic of protection. At the same time, the existing

dilemma of prompt copyright protection in prompt

trading is also clarified.

2.1 The Nature, Classification, and

Characteristics of Prompts

In the field of natural language processing, a prompt

usually refers to an instruction or a set of instructions

for guiding a generative AI to produce desired output

content, which can be in form of natural language and

other forms such as images, with characteristics such

as intangibility, replicability, functionality, as well as

structural hierarchy and composability (Huang &

Rust, 2024 … Schulhoff et al, 2025).

According to application scenarios and functional

differences, prompts can be broadly classified into six

categories: input semantics, output customization,

error identification, prompt improvement,

interaction, and context control; according to the

usage and complexity, prompts can be classified into

general-purpose prompts, application-specific

prompts (e.g., drawing, code generation, etc.), and

nested high-level prompt types. The structure of

prompts usually includes elements such as

information expression, functional instructions, and

creative structure, aiming to guide the model to

generate outputs that meet expectations (White et al,

2023).

Professional, high-quality prompts hold great

potential for their completeness and excellent output

orientation, and users can dominate the generative AI

with these high-quality prompts to do the expected or

even exceed the expected work for the users at the

least possible debugging cost. Prompts are

characterized by two features in terms of value

realization: first, their value is dependent on model

capabilities rather than direct presentation; second,

they are vulnerable to reverse engineering and

cracking after trading, leading to market price

collapse (Wyk et al, 2023).

2.2 The Compatibility Between Prompt

Protection and the Legislative

Purpose and Logic of Copyright

Law

As a property right system to stimulate innovation,

function of "encouraging creativity", under which

"creativity" as an important concept dominates a

series of legal rules (Zhang, 2025). In addition, taking

China's copyright law as an example, the legislative

basis and purpose of Article 1 of the Copyright Law

of the People's Republic of China reflect this core

value and institutional function, and its constitutional

basis lies in the protection of citizens' fundamental

rights, namely, freedom of expression and freedom of

creativity, which also makes it clear that works

protected by the copyright law can only be the

intellectual achievements created by natural persons

(Zhang, 2025).

Due to the varying needs and the absence of a

fixed path or template, prompt creation lacks a one-

to-one correspondence between the prompt and the

intended output. As a result, similar outputs can be

produced by different prompts. This characteristic

grants authors a degree of creative autonomy.

With respect to authorship, although the unique

nature of prompts often necessitates the use of

artificial intelligence—whether for feedback,

optimization, or even direct generation—prompt

creation may involve varying degrees of human and

AI involvement. Nevertheless, prompts that possess

sufficient economic value are, in essence,

predominantly human-directed.

Therefore, prompt engineering as an emerging

discipline, the inclusion of prompts in the scope of

vitality but also incentivize creation and establish a

healthy market based on the exclusive protection of

authors, and it also promotes the development of

artificial intelligence technology.

2.3 The Similarities Between Prompts

and Code in Copyright Protection

Among the types of intangibles traditionally

protected by law, codes and prompts have the most

similarity and referability in terms of their

manifestation and nature, and the feasibility of

prompt copyright protection can be further explored

by analyzing the protection of codes in legislation and

judicial practice.

Firstly, both prompts and code are functionally

oriented artifacts, with authors arranging words to

essentially fulfill a function. Prompts guide the AI

through natural language in a function-like

relationship to generate the intended output, and

similarly, codes are functional tools. Notably, their

functionality is protected through patents in some

countries, such as the United States (Zhao, 2010).

Secondly, the similarity between prompts and

codes in terms of expression form and creation

process is reflected in the fact that both of them are

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

560

textual expressions. The composition of prompts as

natural language instructions is itself a sequential

combination of words, making it possible to become

an original text; while codes are likewise literal forms

of expression, i.e., readable sequences of symbols

(Schulhoff et al, 2025). The U.S. Federal Court of

Appeals for the Third Circuit in 1983 explicitly

defined both source code and object code as "literary

works", and the core of their copyright protection lies

in their "textual expressiveness" (Geissler, 2015).

2.4 The Dilemma of Rights Protection

for Prompts

The realization of the property rights and interests of

AI prompts as commodities in practice has been

greatly impeded by the fact that, due to their special

nature of existing in plain text form, they are

extremely easy to be copied, disseminated, and

tampered with, which seriously undermines their

original market value and trading potential. In the

absence of an exclusive protection mechanism, a

prompt is in the public domain once it has been sold,

making it difficult for authors to control their

subsequent use and circulation, leading to a

downward spiral in market prices and dispersed

revenues, and hampering the normal formation and

sustained development of the market for prompts

(Wyk et al, 2023).

3 ASSESSMENT PATHWAY FOR

THE COPYRIGHTABILITY OF

PROMPTS

3.1 Textual Composition and

Originality of Prompts

In the use of natural language models, prompts are

mainly presented in textual form, and scholars have

affirmed to some extent that "the 'user input' made to

an AI may itself constitute a textual work" when

discussing copyright over AI-generated objects

(Wang, 2024).

3.1.1 Functional Collection of Text

The essence of a prompt lies in a collection of words

that form natural language instructions, with its

originality evident in the systematic arrangement and

functional design of linguistic symbols. Since

"copyright law only protects expressions that

demonstrate originality, and excludes practical

elements such as methods of operation, technical

solutions, and functional applications from its

protective scope", the functionality of prompts—

viewed as text-based instructions—cannot be

afforded copyright protection (Wang, 2021).

In many cases, merely transmitting work

instructions to a generative AI system is insufficient

to satisfy the originality criterion for copyright

protection. Even when the desired outcome is

meticulously delineated, the prompt still cannot be

deemed copyrightable due to its technical

functionality and inherent utility (Verch, 2024).

However, the existence of the prompt as a tool

governing artificial intelligence and its capacity to

fulfill a specific function cannot be disregarded. This

does not, in and of itself, preclude the possibility of

protection under copyright law. Therefore, while

acknowledging its functionality, further exploration

of the intricacy of its text and the substance of its

original expression is necessary to elevate it to a level

where it can be safeguarded by copyright legislation.

3.1.2 Collection of Text with Original

Expressiveness

Unlike other forms of artificial intelligence that

operate based on predetermined paths and fixed

procedures, generative AI introduces an element of

unpredictability in its outputs. This unpredictability

elevates the significance of prompts within human-AI

interactions and has contributed to their increasing

complexity and specialization. As a result, prompts

may embody a certain level of creativity and

expression, potentially qualifying as literary

expressions with originality (Mazzi, 2024).

Accordingly, in case-by-case evaluations, where a

prompt demonstrates a sufficient degree of

originality, it may be recognized as a written work

under copyright law and thereby qualify for legal

protection.

Under the originality standards of different

jurisdictions, the European Union tends to emphasize

whether a prompt demonstrates sufficient creativity

and reflects the personal imprint of a human author

(Mazzi, 2024). In Infopaq International A/S v. Danske

Dagblades Forening, the Court of Justice of the

European Union (CJEU) held that even text

fragments as short as 11 words may qualify as

protectable works, provided they embody the

intellectual choices and expression of the author.

Under the United States standard, originality

requires more than mere labor; prompts must exhibit

a minimal degree of creativity, such as through the

inclusion of unique, imaginative, or innovative

elements (Mazzi, 2024). In Feist Publications, Inc. v.

Copyrightability of Artiﬁcial Intelligence Prompts

561

Rural Telephone Service Co., the U.S. Supreme Court

held that “a modicum of creativity” is the threshold

for copyright protection. As long as the author makes

minimal creative choices—such as in the selection,

coordination, or arrangement of content—even

simple expressions may merit protection (Wang,

2021).

In conclusion, comparative legal standards

regarding the copyrightability of short texts and

works embodying a “minimal degree of intellectual

creativity” may serve as valuable references in

assessing the copyrightability of prompts, which

often take the form of short texts containing both

functional and original expression.

3.2 The Authorial Subject of Prompts

Even if a prompt qualifies as an “original expression,”

it remains necessary to determine the legal authorship

of the work. Due to the unique characteristics of

prompts, their authors may involve hybrid authorship.

This section examines the copyrightability of prompts

under various types of creative authorship structures.

3.2.1 Human-Created or AI-Generated

Prompts

laborers. "Since only human beings can understand

and utilize the incentives of the copyright law, only

the results of human creations can be protected by the

cannot directly determine AI-generated content based

on their free will, thus excluding AI developers and

users from the subjective scope (Wang, 2023).

In other words, to date, in the general practice of

States, copyright protection can only be granted to

works whose authors are "human beings from

beginning to end and only human beings". Therefore,

it is difficult to recognize AI-generated works as the

subject of copyright protection. Therefore, the

copyrightability of a prompt is based on the fact that

the prompt constitutes an original expression and is

created by a human being based on his or her free will.

3.2.2 Human-AI Collaborative Prompts

In the discussion of copyright attribution of AI-

generated objects, it has been argued that since the

subject of its creation is not a human being, and a

human being is unable to decide on the generated

object by his/her free will, and therefore it is difficult

for a human being to obtain copyright even if he or

she has put in a certain degree of labor in the human-

computer interaction.

However, in the process of prompt creation,

interaction with generative AI is virtually inevitable,

whether for testing, generating, or optimizing

prompts. A blanket exclusion of copyright protection

for prompts involving AI intervention could pose

significant risks to the development of the industry.

Therefore, in assessing the copyrightability of

prompts, it is advisable to adopt a more flexible

approach to the human–machine collaboration

paradigm within a clearly defined institutional

framework. For instance, the traditional theory of

authorship under copyright law may be reconsidered

and restructured to accommodate a new category of

collaborative works through a “dual-subject model of

creation.” Under this model, works designed through

human–machine collaboration could be recognized as

joint creations of both human and machine authors

(Wu, 2024).

3.3 Intellectual Labor in the Process of

Creating Prompts

Advanced and high-quality prompts are characterized

by structural complexity, often comprising multiple

layers such as basic directives, creative arrangements,

and system-level architecture. Furthermore, due to

the stochastic nature of large language models, where

identical prompts may yield varying outputs across

multiple iterations, it becomes necessary for prompt

designers to engage in extensive testing and

continuous refinement to maintain output quality in

the face of such unpredictability (Wyk et al, 2023 &

Zamfirescu-Pereira et al, 2023). In domain-specific

contexts, such as medicine, prompt engineering

requires the creator to possess substantial subject-

matter expertise. By contrast, non-experts often adopt

ad hoc or opportunistic approaches to prompt design,

which tend to lack systematicity and are prone to

overgeneralization or excessive reliance on personal

interaction experience (Zamfirescu-Pereira et al,

2023). This evidences that prompt creation transcends

mere instruction-giving or mechanical compilation. It

is not simply the result of “sweat-of-the-brow” labor,

but rather a form of intellectual creation that may

qualify for copyright protection.

4 RECOMMENDATIONS FOR

THE LEGAL PROTECTION

PATHWAYS OF PROMPTS

Admittedly, the intrinsic characteristics of prompts,

coupled with technological limitations, pose

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

562

significant challenges to the realization of their

property rights. Nevertheless, legal regulation can

play a pivotal role in promoting the recognition and

protection of prompts across other domains. That

said, such regulation must be carefully calibrated to a

minimal threshold. Excessive legal protection risks

fostering monopolistic practices through the abuse of

rights, thereby undermining technological innovation

and impeding the broader development of AI-driven

industries.

4.1 Recommendations for the Legal

Protection Pathways of Prompts

In view of the rapid pace of technological

advancement, it is more feasible to expand the scope

of judicial interpretation to clarify whether prompts

can be considered objects of copyright protection and

to establish the corresponding criteria for such

recognition. This would help prevent the inclusion of

prompts that are overly simplistic, highly functional,

or fail to meet authorship requirements from falling

within the scope of protection.

Some countries have extended the applicability of

existing copyright laws to address the

copyrightability of AI-generated content. For

instance, the UK Intellectual Property Office (IPO),

in its consultation paper “Artificial Intelligence and

Intellectual Property”, stated that AI-generated

content may be protected under the current legal

framework. It also expressed its intention to remain

engaged at the international level and to revise,

replace, or repeal relevant protective clauses as

necessary (Wu, 2024).

Currently, there is no judicial precedent explicitly

addressing the copyrightability of prompts. In

practice, determinations rely heavily on judicial

discretion in individual cases, due to the absence of

uniform standards. The regulation of prompts could

draw on existing approaches to AI-generated works,

thereby enhancing legal flexibility without the

immediate need for dedicated legislation.

4.2 Establishment of a New Paradigm

for Human-AI Collaboration

Given the inseparability of prompt creation from

generative AI, a new paradigm may be developed—

one that emphasizes the substantial involvement of

human authors in the creative process and adopts a

dual-ownership framework. This approach can draw

upon existing scholarship on models of human-

machine collaboration in the context of the

copyrightability of AI-generated outputs (Wu, 2024).

4.3 Strengthening Industry and Market

Self-Regulation of Prompt Use

Various types of generative AI models and prompt

trading platforms should actively formulate industry

standards. These standards should clarify the rights

and obligations of generative AI service providers,

service users, prompt authors, and prompt consumers.

They should also provide specific protection for

prompts involving copyright and regulate behaviours

with infringement potential. This will promote the

reasonable distribution of rights and protection in the

industry through self-regulatory mechanisms. This

promotes the reasonable distribution of rights and

protection in the industry, avoids excessive legal

intervention through independent industry regulation,

and further standardises the market.

5 CONCLUSION

With the rapid development and popularisation of

generative artificial intelligence in natural language

processing, legal disputes triggered by it have become

an urgent and unavoidable response to the reality of

the problem. The importance of prompts as core

inputs in generative artificial intelligence operations

has grown, giving rise to a significant prompts trading

market. This has prompted reflection on copyright

protection for prompts.

Although prompts have significant functional

attributes that prevent them from being included in

the traditional scope of copyright protection to a

certain extent, it should also be recognised that, under

the development of artificial intelligence, prompts are

becoming more complex and professional texts that

need to be arranged and debugged by authors.

Therefore, some prompts can have original

expressions, and the possibility and reasonableness of

obtaining copyright protection for such prompts

should be recognised. The possibility and

reasonableness of obtaining copyright protection for

such prompts should be recognised. Furthermore,

while prompt creation relies on artificial intelligence

for feedback and optimization, human intervention

and intellectual labour are still substantial, so the new

paradigm of human-computer collaboration can be

standardised and regulated within certain limits.

Furthermore, given that prompt creation often

requires specialised knowledge and multiple rounds

of debugging, it maps to a higher degree of

intellectual work, and therefore has economic value.

Although the current international judicial

practice does not legally recognise or protect the

Copyrightability of Artiﬁcial Intelligence Prompts

563

rights of prompts, action taken by countries such as

the United Kingdom can be taken as reference,

namely expanding the scope of application of the law

through judicial interpretation, to clarify the uniform

determination standard of the copyrightability of

prompts. Furthermore, a novel paradigm of human-

computer collaboration should be established, and the

human-computer collaboration model should be

strengthened. Additionally, the industry's self-

regulation should be enhanced to further regulate the

market.

In conclusion, while there are certain challenges

in determining appropriate and clear copyright

protection for AI prompts, establishing their

copyrightability protects the property rights and

interests of prompt authors. It also provides a

systematic basis for protecting the interests of those

trading in prompts and consumers and empowers the

sustainable development of AI technology.

Furthermore, it protects scientific and technological

innovation, as well as industrial upgrading.

REFERENCES

C. J. Geissler, 2015. Oracle America, Inc. v. Google, Inc.,

750 F.3d 1339 (Fed. Cir. 2014). Cert. denied: Ideas,

methods, and expression - whose innovation is

protected? Cath. U. J.L. & Tech. 24, 264.

F. Mazzi, 2024. Authorship in artificial intelligence-

generated works: Exploring originality in text prompts

and artificial intelligence outputs through philosophical

foundations of copyright and collage protection. J.

World Intellect. Prop., 27, 414.

H. Wu, 2024. On the copyrightability of AI-generated

content: Practice, legal theories, and institutional

frameworks. China Law Rev., 3, 127.

protection by watermark injection and verification.

arXiv: 2308.02816, 3.

J. D. Zamfirescu-Pereira et al., 2023. Why Johnny can't

prompt: How non-AI experts try (and fail) to design

LLM prompts. Proc. CHI Conf. Human Factors

Comput. Syst., ACM, 2.

J. White et al., 2023. A prompt pattern catalog to enhance

prompt engineering with ChatGPT. arXiv: 2302.11382.

L. Zhang, 2025. Reanalyzing the legal nature of AIGC

works from the perspective of legal philosophy.

Intellect. Prop. 3, 12.

L. Zhao, 2010. Conflicts of legal philosophy and technical

dilemmas in patent protection of software source code.

Legal Syst. Soc., 8.

M. H. Huang, R.T. Rust, 2024. Automating creativity.

arXiv: 2405.06915, 16.

M.A. van Wyk et al., 2023. Protect your prompts: Protocols

for IP protection in LLM applications.

arXiv:2306.06297, 1.

P. Liu et al., 2023. Pre-train, prompt, and predict: A

systematic survey of prompting methods in natural

language processing. ACM Comput. Surv. 55, 195.

Q. Wang, 2021. Textbook of intellectual property law (7th

ed.) (Renmin Univ. China Press, 2021), 92

classification of AI-generated content. Trib. Political

Sci. Law, 41, 16.

Q. Wang, 2024. A Third Inquiry into the Legal Status of AI-

Generated Content under Copyright Law, Law Bus.

Stud., 41, 182.

S. Schulhoff et al., 2025. The prompt report: A systematic

survey of prompt engineering techniques.

arXiv:2406.06608, 5.

U. Verch, 2024. Per Prompt zum Plagiat?: Rechtssicheres

Publizieren von KI-generierten Inhalten, API Mag., 5,

9–10.

X. Amatriain, 2024. Prompt design and engineering:

Introduction and advanced methods. arXiv:

2401.14423, 1.

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

564