Digital Art Fraud in India: NFT Misappropriation, AI Scraping, Style Mimicry and Print-on-Demand Theft

Introduction : Digital art has expanded the creative economy by allowing artists to produce, publish and monetise work without traditional galleries or physical intermediaries. The same features that make digital art accessible also make it vulnerable to rapid copying. A work can be downloaded, altered, uploaded to a marketplace, used to train an artificial-intelligence system or printed on commercial products within minutes.

Four forms of misuse are particularly significant in India: unauthorised NFT minting, AI scraping, style mimicry and print-on-demand theft. They do not create identical legal problems. An unauthorised NFT may involve direct reproduction and false provenance. AI scraping may involve mass copying for training, with difficult questions about fair dealing and dataset use. Style mimicry may not infringe copyright if it copies only an artistic style rather than protected expression. Print-on-demand theft commonly involves reproduction, commercial distribution and possible trademark misuse.

The legal framework is therefore distributed across the Copyright Act, 1957, the Trade Marks Act, 1999, the Information Technology Act, 2000, contract law and intermediary principles. India does not yet have a comprehensive statute specifically designed for digital art fraud or generative-AI copying. Effective protection consequently depends on applying existing rights carefully and combining legal action with technical and contractual safeguards.

Unauthorised NFT Minting

An NFT is a blockchain-based token that identifies or represents a digital asset. It may be associated with an artwork, animation, photograph, music file or video. The token records ownership or transfer of the token itself, but it does not automatically transfer copyright in the underlying work.

This distinction is critical. Under Section 14 of the Copyright Act, the copyright owner controls reproduction, issuing copies, communication to the public, adaptation and other forms of exploitation. Minting an NFT linked to an artist’s work may require copying the work to a marketplace, server or metadata file. Selling the token may also communicate the work to the public. If the minter lacks permission, Section 51 may be attracted.negd.gov+1

The purchaser of an NFT also does not automatically acquire copyright. A copyright assignment must be in writing and must identify the rights assigned, duration and territorial extent under Section 19. Unless the transaction expressly provides otherwise, the buyer may obtain only the token and a limited licence to display the associated work. 

Unauthorised minting may involve additional wrongs. If the minter falsely represents that they created or own the work, the conduct may support passing off, fraud or misrepresentation claims. If the artist’s name or signature is removed or altered, moral-rights concerns may arise under Section 57 of the Copyright Act. 

A marketplace may face separate exposure if it knowingly promotes stolen art, ignores specific notices or presents the seller as an authenticated creator. Its liability will depend on its role, knowledge, platform terms and applicable intermediary protections. Technical decentralisation does not eliminate copyright infringement because the blockchain records a transaction, not lawful authority.

AI Scraping and Training Data

AI scraping involves collecting digital artworks from websites, social-media pages, online galleries or image repositories and using them to train or improve an artificial-intelligence model. The process may involve downloading, storing, converting and analysing the works.

The Copyright Act protects original artistic works under Section 13, and Section 14 gives the owner exclusive reproduction rights. The initial copying involved in scraping can therefore raise infringement concerns. The fact that the system converts images into data representations does not automatically answer whether the underlying reproduction was authorised. 

Section 52 contains fair-dealing exceptions, including private or personal use and research, but it does not create a broad, express text-and-data-mining exception for commercial generative-AI systems. A developer may argue that training is transformative research and that the model does not distribute the original artworks. The rights holder may respond that the training process involved systematic copying and that the model competes with the artist’s licensing market. 

The legal analysis should separate training from output. Even if a court considers some training activity permissible, an AI system that generates a substantially similar copy, reproduces a protected work or enables users to obtain memorised images may create a separate infringement risk. The output must be evaluated independently.

Indian law also protects compilations and databases where originality exists. A curated dataset may therefore involve rights in both individual artworks and the selection or arrangement of the collection. Contractual restrictions on scraping may apply even where copyright protection in individual images is uncertain.

The safest model for commercial developers is permission-based licensing. Agreements should identify the works covered, permit ingestion and storage, regulate model training and fine-tuning, address generated outputs, provide attribution or remuneration where agreed and establish opt-out or takedown procedures. Developers should also maintain provenance records and exclude material whose licensing status is unclear.

Style Mimicry

Style mimicry presents a more difficult problem because copyright generally protects expression, not artistic style, ideas, methods or aesthetic concepts. A visual style may include colour palettes, brush techniques, compositional tendencies or recurring themes. These features may be commercially distinctive but may not, by themselves, constitute a protected work.

The legal position changes if the imitation reproduces substantial expressive elements of a particular artwork. An AI-generated image may be described as being “in the style of” an artist, yet still infringe if it copies a recognisable composition, characters, pose, arrangement or other protected expression. The question is not whether the image feels similar in a broad aesthetic sense. It is whether protected expressive choices have been appropriated.

Style mimicry may also engage trademark and passing-off principles. If the image is marketed as an artist’s work, falsely suggests endorsement or uses the artist’s name to attract customers, the artist may challenge the misrepresentation. An artist’s name, logo or signature may also function as a trademark.

Personality and moral-rights arguments may arise where the use falsely attributes a work to the artist or harms the artist’s reputation. Section 57 protects certain authorial interests, including the right to claim authorship and object to distortion or mutilation prejudicial to the author’s honour or reputation. 

The difficulty is evidentiary. An artist must distinguish unlawful copying from legitimate influence or independent creation. Evidence may include training prompts, dataset records, repeated reproduction of distinctive features, marketing statements, attribution claims and comparison with identified works.

Print-on-Demand Theft

Print-on-demand platforms allow users to upload designs that are printed on clothing, posters, phone cases, bags and other products only after an order is placed. This reduces inventory costs but makes unauthorised commercial reproduction easy.

Uploading a digital artwork for printing may infringe the copyright owner’s reproduction right. Printing and selling the product may involve issuing copies, commercial distribution and communication to the public. If the design includes a brand logo, character or photograph, trademark and related rights may also be engaged.

The platform’s responsibility depends on its actual role. A service that merely provides technical hosting may rely on intermediary protection if it satisfies statutory conditions and due diligence. A platform that selects designs, promotes them, verifies sellers or profits from known infringing material may face a more demanding analysis.

Artists should preserve evidence of the original work, upload date, registration or publication records, product listings, seller information, orders and sales data. A prompt and specific notice should identify the work, ownership basis, infringing URL and requested action. Vague complaints covering an entire catalogue are less effective than itemised notices.

Available Legal Protections

The Copyright Act provides civil remedies including injunctions, damages, accounts of profits and delivery-up or destruction of infringing copies under Section 55. In appropriate cases, criminal sanctions may apply under Section 63 for knowing infringement or abetment.

Section 65A addresses circumvention of technological protection measures, while Section 65B concerns the protection of rights-management information. These provisions may be relevant where a fraudster removes digital watermarking, creator information or other rights-management data. 

The Trade Marks Act may apply where the artist’s name, logo, signature or brand is used to create false association. Section 29 addresses infringement of registered marks, while Section 27(2) preserves passing off for unregistered goodwill. 

The Information Technology Act may become relevant when fraud includes identity theft, cheating by personation, unauthorised access or misuse of computer resources. Sections 66C and 66D address identity theft and cheating by personation using a computer resource. These provisions do not replace copyright remedies, but they may apply where the digital art fraud is part of a broader deception. 

Mechanisms to Curb Digital Art Fraud

Protection should operate at several levels. Artists should retain layered source files, sketches, working files, metadata and dated exports. Copyright registration is not mandatory, but registration and reliable creation records can strengthen proof of ownership. Watermarks, embedded metadata and cryptographic hashes can support authentication. Licences should define reproduction, display, adaptation, minting, merchandising, AI training, commercial use, attribution and sublicensing.

NFT terms should clearly distinguish token ownership from copyright ownership. Print-on-demand contracts should prohibit uploads without verified rights and require seller indemnities. Marketplaces should use creator verification, provenance checks, repeat-infringer policies and item-specific notice-and-takedown systems. AI developers should maintain dataset documentation, licensing records, opt-out mechanisms and output safeguards. Platforms should preserve transaction and access records so that anonymous actors can be traced where legally possible.

Conclusion

Digital art fraud in India may take the form of unauthorised NFT minting, AI scraping, style mimicry or print-on-demand theft. The legal consequences differ, but the Copyright Act 1957 remains the central framework. Trademark, moral-rights, contract and Information Technology Act provisions may supplement it where the conduct includes false attribution, brand misuse, identity theft or online deception.

NFT ownership does not automatically transfer copyright. AI training is not automatically protected by fair dealing. Style alone may not be copyrightable, but copying expressive elements or falsely invoking an artist’s identity can create liability. Print-on-demand reproduction is ordinarily high-risk where the uploader lacks permission.

The practical solution is a combination of documentation, licensing, technical monitoring, provenance verification, platform cooperation and timely legal enforcement. Digital art can be commercially innovative without being legally unregulated. The essential principle remains that technology may change how a work is copied, sold or transformed, but it does not remove the creator’s rights in the original expression.

Author: Amrita Pradhan in case of any queries please contact/write back to us via email to content@khuranaandkhurana.com or at  Khurana & Khurana, Advocates and IP Attorney.

References

1. Copyright Act, 1957, Section 13.

2. Copyright Act, 1957, Section 14.

3. Copyright Act, 1957, Section 51.

4. Copyright Act, 1957, Section 19.

5. Copyright Act, 1957, Section 57.

6. Copyright Act, 1957, Section 52(1)(a).

7. Copyright Act, 1957, Section 55.

8. Copyright Act, 1957, Section 63.

9. Copyright Act, 1957, Section 65A.

10. Copyright Act, 1957, Section 65B.

11. Trade Marks Act, 1999, Section 27(2).

12. Trade Marks Act, 1999, Section 29.

13. Trade Marks Act, 1999, Section 30.

14. Information Technology, Act 2000, Section(s) 66C and 66D.

15. Press Information Bureau, ‘Existing IPR Regime Well-Equipped to Protect AI-Generated Works’ (9 February 2024) https://www.pib.gov.in/PressReleasePage.aspx?PRID=2004715

16. Nishith Desai Associates, Tracking NFTs from Code to Court: Legal Considerations and Disputes (2024).

17. World Intellectual Property Organization, ‘Copyright and Artificial Intelligence’ https://www.wipo.int