Attribution requirements for AI assisted work

Introduction

Attribution requirements for AI-assisted work refer to the legal, academic and ethical obligations concerning the identification, acknowledgement and disclosure of artificial intelligence tools used in creating written, artistic, scientific, academic or professional content. With the rapid development of generative artificial intelligence, individuals increasingly use AI systems to prepare research papers, articles, reports, software, photographs, designs and other creative materials.

AI-assisted work creates important questions concerning authorship, originality, copyright ownership, plagiarism, academic integrity and intellectual property rights. One major concern is whether a person must disclose the use of AI when submitting or publishing work and whether AI systems can be recognised as authors or inventors.

Attribution requirements differ according to the nature of the work, institutional policies, publishing standards and applicable copyright legislation. In India, the Copyright Act, 1957, the Information Technology Act, 2000, and applicable institutional regulations provide relevant legal foundations. Internationally, copyright principles and emerging AI governance frameworks influence the development of attribution standards.

Meaning and scope of AI attribution

AI attribution means acknowledging the contribution of artificial intelligence tools in producing or modifying a particular work.

It involves identifying whether AI was used for generating content, editing language, preparing illustrations, analysing information, developing software or assisting with research.

Attribution requirements generally cover the following areas:

AI disclosure: Informing readers, publishers or institutions that AI tools were used.

Human authorship: Identifying the individual responsible for the intellectual contribution and final work.

Copyright ownership: Determining who legally owns the resulting creative material.

Source acknowledgement: Identifying original sources used to support AI-generated information.

Academic integrity: Preventing undisclosed AI assistance from violating institutional rules.

Research transparency: Explaining the role of AI in research methodology and scientific publications.

Accountability: Ensuring that human authors remain responsible for accuracy and originality.

Attribution does not automatically establish copyright ownership. Similarly, mentioning an AI tool does not necessarily establish that the resulting work is original or legally protected.

International legal framework governing AI attribution

1. Berne Convention for the Protection of Literary and Artistic Works, 1886

The Berne Convention establishes international principles concerning copyright protection.

Article 6bis recognises authors' moral rights, including the right to claim authorship and object to certain distortions or modifications of a work that may prejudice the author's honour or reputation.

The Convention is relevant to AI-assisted work because it provides an international foundation for recognising human authorship and protecting the integrity of creative works.

However, it does not establish a comprehensive international disclosure requirement for the use of generative AI.

2. Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), 1994

The TRIPS Agreement establishes minimum standards for intellectual property protection among World Trade Organization members.

Article 9(2) provides that copyright protection extends to expressions and not ideas, procedures, methods of operation or mathematical concepts as such.

This distinction is important when evaluating AI-assisted works because copyright protection generally concerns original human expression rather than abstract ideas or automatically generated information.

3. World Intellectual Property Organization (WIPO) framework

WIPO has examined the relationship between artificial intelligence and intellectual property through its discussions on intellectual property policy and generative AI.

Important issues include:

Copyright ownership of AI-generated works.

Recognition of human creative contributions.

Disclosure of AI involvement.

Protection of training data and copyrighted material.

Attribution of original creators.

Responsibility for copyright infringement.

WIPO's discussions demonstrate the need for international cooperation, although they do not create a universal mandatory AI attribution rule.

4. UNESCO Recommendation on the Ethics of Artificial Intelligence, 2021

UNESCO's Recommendation establishes ethical principles for responsible AI development and use.

It emphasises transparency, accountability, human oversight, fairness and respect for human rights.

These principles support appropriate disclosure of AI involvement in academic, professional and creative activities.

However, the Recommendation is a non-binding international instrument rather than a directly enforceable copyright law.

5. European Union Artificial Intelligence Act, 2024

The EU AI Act establishes transparency obligations for certain AI-generated or manipulated content.

Article 50 contains transparency requirements concerning certain AI interactions and synthetic audio, image, video and text content, subject to the applicable conditions and exceptions.

These provisions are relevant to identifying AI-generated material and preventing deception.

The Act does not establish a universal requirement that every person must acknowledge every use of AI in every written work.

Legal framework governing AI attribution in India

1. Copyright Act, 1957

The Copyright Act is the principal legislation governing copyright protection in India.

Section 2(d)(vi) defines the author of a computer-generated literary, dramatic, musical or artistic work as the person who causes the work to be created.

This provision is particularly relevant to AI-assisted content because it raises questions concerning the identification of the human person responsible for causing a computer-generated work to be created.

Section 13 establishes the categories of works in which copyright subsists, subject to statutory requirements.

Section 14 defines the exclusive rights of copyright owners.

Section 17 identifies the first owner of copyright, subject to statutory exceptions.

Section 57 recognises certain authors' special rights, including the right to claim authorship.

However, the Act does not provide a complete or explicit framework for determining authorship and ownership in every situation involving modern generative AI.

2. Information Technology Act, 2000

The Information Technology Act provides the legal framework for electronic records, digital transactions and certain forms of cyber-related misconduct.

Sections 43 and 66 address specified forms of unauthorised access and computer-related offences.

Section 79 establishes conditional safe-harbour protection for eligible intermediaries, subject to statutory requirements.

These provisions may become relevant where AI-assisted work involves unauthorised access, digital content distribution or intermediary liability. However, the Act does not independently establish general AI attribution requirements.

3. National Education Policy, 2020

The National Education Policy promotes academic integrity, critical thinking, research quality and responsible educational practices.

Educational institutions may develop policies governing the use of AI tools in assignments, examinations, dissertations and research publications.

Students and researchers may be required to disclose AI assistance according to institutional rules.

4. University and research institution regulations

Universities and research institutions establish academic integrity policies concerning plagiarism, originality and proper acknowledgement of research assistance.

AI-generated content may create academic misconduct issues where students submit work in violation of institutional requirements or misrepresent the origin of their contributions.

The University Grants Commission's academic integrity framework, including the UGC Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions Regulations, 2018, is relevant to originality and plagiarism assessment.

However, AI-generated content is not automatically plagiarism. Its treatment depends on the applicable academic rules, the originality of the material and the manner in which it is presented.

Major legal issues concerning AI attribution

1. Identification of human authorship

One of the most important questions is whether an AI system can be recognised as an author.

Under existing copyright frameworks, legal authorship generally remains connected to a legally recognised human or other qualifying legal person, depending on the jurisdiction.

In India, Section 2(d)(vi) of the Copyright Act creates a specific provision concerning computer-generated works.

The challenge is determining who caused the work to be created when several people, software developers and AI systems contribute to the final output.

2. Copyright ownership of AI-assisted work

AI-assisted work may contain both human-created and machine-generated elements.

For example, a researcher may develop an original argument, use AI for language correction and independently revise the final manuscript.

In such a situation, the researcher's original human expression may qualify for copyright protection, subject to the statutory requirements.

However, purely AI-generated material raises more difficult questions regarding originality and ownership.

3. Academic integrity and plagiarism

Students and researchers may use AI to prepare essays, dissertations, research summaries and scientific publications.

Undisclosed AI assistance may violate academic policies where disclosure is mandatory.

A separate concern arises when AI-generated material reproduces copyrighted sources without appropriate acknowledgement.

Institutions should distinguish between legitimate AI assistance, plagiarism, fabricated references and misrepresentation of research contributions.

4. Attribution of training data

Generative AI systems may be trained using large collections of books, photographs, articles, software and other copyrighted materials.

This creates legal questions concerning the recognition of original creators and the lawful use of protected works.

The use of copyrighted material in AI training and the obligation to attribute training sources are separate legal questions. Attribution alone does not automatically make an otherwise unauthorised use lawful.

5. Accountability for inaccurate information

AI systems may produce inaccurate statements, fabricated references and misleading research findings.

Human authors and publishers must therefore maintain appropriate verification procedures.

Attribution should identify the role of AI without transferring responsibility for the final work entirely to the technology provider.

Important case laws and judicial principles

There are no widely recognised Indian judicial decisions directly establishing a comprehensive attribution framework for generative AI-assisted work. However, the following decisions provide important principles concerning authorship, originality and intellectual property.

1. Eastern Book Company v. D.B. Modak (2008)

The Supreme Court of India examined copyright protection in judicial decisions and the originality required for copyright protection.

The Court adopted a standard requiring a sufficient degree of skill and judgment, rather than treating mere labour as sufficient for originality.

Legal principle: Copyright protection requires original intellectual contribution and cannot be established merely through mechanical effort.

Relevance: The judgment is important when assessing whether human modifications, selections or arrangements of AI-generated material contain sufficient original human intellectual contribution.

2. University of London Press Ltd. v. University Tutorial Press Ltd. (1916)

The English court considered copyright protection in examination papers and explained that originality does not require novelty or inventive genius.

A work may be original where it originates from its author and involves sufficient independent intellectual effort.

Legal principle: Originality concerns the independent creation of expression rather than whether the underlying subject matter is entirely new.

Relevance: This principle helps explain why human contributions to AI-assisted writing may qualify for copyright protection even when AI tools are used during preparation.

3. Feist Publications, Inc. v. Rural Telephone Service Co. (1991)

The United States Supreme Court held that copyright protection requires originality involving independent creation and at least a minimal degree of creativity.

The Court rejected the proposition that labour or effort alone automatically creates copyright protection.

Legal principle: Originality requires independent creation and a minimum level of creativity.

Relevance: This judgment provides comparative guidance for evaluating the copyright status of AI-assisted compilations, databases and creative works.

4. Naruto v. Slater (2018)

The United States Court of Appeals for the Ninth Circuit considered whether a monkey could claim copyright ownership over photographs it had taken.

The court concluded that the Copyright Act did not authorise the animal to bring the copyright claim.

Legal principle: Copyright authorship and enforcement depend on the legal requirements of the relevant copyright legislation.

Relevance: Although the case concerned an animal rather than artificial intelligence, it is frequently discussed in debates concerning whether non-human entities can qualify as copyright authors.

5. Thaler v. Perlmutter (2023)

A United States federal court examined a copyright application for an image generated autonomously by an artificial intelligence system.

The court upheld the position that human authorship is required under the United States Copyright Act.

Legal principle: A work created entirely by AI without human authorship does not qualify for copyright protection under the relevant United States framework.

Relevance: The judgment is important for international discussions concerning the ownership and attribution of purely AI-generated artistic material.

6. Thaler v. Vidal (2022)

The United States Court of Appeals for the Federal Circuit considered whether an artificial intelligence system could be identified as an inventor under patent law.

The court held that an inventor under the applicable United States Patent Act must be a natural person.

Legal principle: Inventorship under the relevant patent legislation requires a natural person.

Relevance: The decision demonstrates the importance of distinguishing between AI-assisted human innovation and inventions attributed exclusively to AI systems.

Recommended attribution standards for AI-assisted work

A transparent attribution framework should identify the nature and extent of AI involvement.

The following practices are useful:

Identify the AI tool used in preparing the work.

Explain whether AI was used for drafting, editing, translation, research or image generation.

Identify the human author responsible for the final content.

Verify all facts, citations and references independently.

Acknowledge original sources used in the work.

Follow institutional and publisher-specific AI disclosure policies.

Avoid identifying AI systems as human authors or inventors where the applicable law does not recognise them.

Maintain records of substantial AI contributions where required.

Ensure that AI-generated content does not reproduce protected material unlawfully.

Disclose material AI assistance in research publications where the relevant journal requires it.

Conclusion

Attribution requirements for AI-assisted work are becoming increasingly important in copyright law, academic integrity, scientific research and professional publishing. The widespread use of generative AI has created new legal questions concerning human authorship, originality, ownership, disclosure and accountability.

In India, the Copyright Act, 1957, particularly Section 2(d)(vi), provides an important statutory foundation for understanding authorship in computer-generated works. International frameworks, including the Berne Convention, TRIPS Agreement and UNESCO Recommendation on the Ethics of Artificial Intelligence, provide additional guidance concerning copyright and responsible AI use.

Judicial decisions such as Eastern Book Company v. D.B. Modak, Feist Publications v. Rural Telephone Service, Naruto v. Slater and Thaler v. Perlmutter help explain the relationship between originality, human authorship and intellectual property protection.

Ultimately, an effective AI attribution framework should promote transparency, recognise genuine human intellectual contributions, protect original creators and ensure accountability for the final work. It should also distinguish between legitimate AI assistance and the misrepresentation of machine-generated material as independently produced human work.