Author: Smriti Ranjan, Maharishi Markandeshwar (Deemed to be University)
Abstract:
Artificial Intelligence (AI) has transformed modern commercial practices, including the creation, marketing and protection of trademarks. Generative AI can independently produce brand names, logos, slogans, advertisements and other forms of brand identity. While this technology promotes innovation and efficiency, it simultaneously creates novel challenges for trademark law, particularly concerning infringement, deceptive similarity, consumer confusion, passing off, attribution of liability and protection against AI-generated brand impersonation.
The Indian trademark regime, primarily governed by the Trade Marks Act, 1999, was developed before the widespread emergence of generative AI. Although existing provisions are sufficiently broad to address several forms of AI-assisted infringement, significant questions remain regarding liability where the infringing content is generated autonomously by an AI system.
This article examines the emerging intersection between AI and trademark law in India and evaluates whether the existing legal framework is capable of addressing AI-driven trademark violations. It argues that technological innovation must be accompanied by stronger human oversight, effective enforcement mechanisms and appropriate legal clarification.
Keywords: Artificial Intelligence, Trademark, Infringement, Passing Off, Generative AI, Consumer Confusion, Deceptive Similarity.
To the Point:
The growth of AI has fundamentally changed the way businesses create and protect their brands. AI tools can generate hundreds of potential trademarks within seconds and can also analyse existing marks for similarities. However, the same technology can be exploited to create deceptively similar trademarks, counterfeit branding, fake advertisements and AI-generated content falsely suggesting an association with an established brand.
Under Section 29 of the Trade Marks Act, 1999, unauthoriseduse of an identical or deceptively similar registered trademark may constitute infringement when the statutory requirements are satisfied. Section 29 also provides enhanced protection to well-known marks in appropriate circumstances.
The principal difficulty in the AI context is attribution of liability. An AI system itself does not possess legal personality. Therefore, when an AI-generated advertisement infringes a trademark, questions arise regarding whether liability should rest upon the user, developer, commercial beneficiary or platform.
Another concern is consumer confusion. AI-powered search engines and shopping assistants increasingly influence purchasing decisions. A consumer may receive an AI-generated recommendation that incorrectly associates one product with another established brand.
AI can nevertheless be used positively for trademark enforcement. Automated monitoring systems can detect unauthorised uses of marks across websites, social media and online marketplaces. Thus, AI functions both as a potential instrument of infringement and an enforcement mechanism.
Use of Legal Jargon:
1. Deceptive Similarity
Section 2(1)(h) of the Trade Marks Act, 1999 defines “deceptively similar” marks. The test essentially concerns whether the similarity between marks is such that it is likely to deceive or cause confusion. In the AI context, deceptive similarity may arise when an AI-generated brand name, logo or advertisement closely resembles an existing trademark.
2. Trademark Infringement
Section 29 of the Trade Marks Act provides the statutory framework for infringement of registered trademarks. AI-generated commercial content may fall within its scope where the statutory requirements concerning use of the mark are satisfied.
3. Passing Off
Passing off is a common-law remedy protecting the goodwill and reputation of an unregistered trademark. The traditional trinity consists of:
Goodwill + Misrepresentation + Damage
AI-generated advertisements or synthetic media falsely representing an association with an established business may potentially give rise to an action for passing off.
4. Well-Known Trademark
Section 2(1)(zg) recognises the concept of a “well-known trademark”. Section 29(4) further provides protection against certain unauthorised uses of well-known registered marks even in relation to dissimilar goods or services. This protection is particularly relevant in the AI environment because generative systems can reproduce famous brands across unrelated commercial contexts.
5. Likelihood of Confusion
Likelihood of confusion is a central principle of trademark law. Courts consider factors such as visual, phonetic and conceptual similarity, the nature of goods and services, purchasing circumstances and the overall impression created upon the relevant consumer. AI-driven recommendation systems may create new forms of consumer confusion by automatically associating unrelated products or businesses.
6. Human Oversight and Attribution
AI does not independently possess legal personality. Therefore, determining the proximate human actor responsible for an AI-generated infringement is essential. The extent of human control, commercial benefit, knowledge and participation may become relevant in determining liability.
The Proof:
The Indian legal framework already provides several principles capable of addressing AI-related trademark disputes.
Section 9 of the Trade Marks Act deals with absolute grounds for refusal of registration, including lack of distinctive character and marks capable of deceiving or confusing the public.
Section 11 provides relative grounds for refusal where a proposed mark conflicts with an earlier trademark, particularly where there exists a likelihood of confusion or association.
Section 29 provides protection against infringement of registered trademarks.
Section 30 recognises limitations on the effect of registered trademarks in specified circumstances.
The legal position demonstrates that the Trade Marks Act is technology-neutral. It does not matter whether an infringing mark was created manually or through an AI system; the crucial question is whether the statutory requirements for infringement or passing off are fulfilled.
However, AI creates an evidentiary challenge. Establishing who generated an infringing output, who instructed the AI, who commercially exploited it and whether the user knew about the existing trademark may require examination of prompts, system logs, metadata, platform records and digital evidence.
Therefore, the existing law may address the substantive infringement, but procedural and attribution-related questions require further development.
Case Laws
1. Kaviraj Pandit Durga Dutt Sharma v. Navaratna Pharmaceutical Laboratories, AIR 1965 SC 980
The Supreme Court distinguished between infringement and passing off and explained that infringement primarily concerns the rights arising from registration of a trademark.
Relevance to AI:
Where AI-generated content uses a registered mark in a manner falling within Section 29, the registered proprietor may rely upon statutory infringement remedies irrespective of whether the infringing material was created manually or through AI.
2. Ruston & Hornsby Ltd. v. Zamindara Engineering Co., AIR 1970 SC 1649
The Supreme Court emphasised the importance of determining whether the impugned mark is deceptively similar to the registered trademark.
Relevance to AI:
AI-generated brand names may be deceptively similar even when the AI system does not intentionally copy the original mark. The legal inquiry remains focused on the similarity and likelihood of confusion.
3. Cadila Health Care Ltd. v. Cadila Pharmaceuticals Ltd., (2001) 5 SCC 73
The Supreme Court laid down important principles concerning deceptive similarity and likelihood of confusion, particularly emphasising the need for a stricter approach where consumer confusion may have serious consequences.
Relevance to AI:
The case provides a useful framework for assessing whether AI-generated brand identities are likely to confuse consumers.
4. Satyam Infoway Ltd. v. Sifynet Solutions Pvt. Ltd., (2004) 6 SCC 145
The Supreme Court recognised that domain names can possess characteristics similar to trademarks and receive protection against passing off.
Relevance to AI:
The decision demonstrates the judiciary’s willingness to apply traditional trademark principles to emerging digital technologies. Similar reasoning may assist courts in addressing AI-generated brand identities and online impersonation.
5. N.R. Dongre v. Whirlpool Corporation, 1996 5 SCC 714
The Supreme Court recognised the protection available to a well-known mark against unauthorised use that could mislead consumers.
Relevance to AI:
AI-generated advertisements or synthetic content using famous brands may exploit the goodwill and reputation attached to well-known trademarks.
6. Yahoo Inc. v. Akash Arora, 1999 SCC OnLine Del 177
The Delhi High Court granted protection against the unauthorised use of a deceptively similar domain name.
Relevance to AI:
The case illustrates the judiciary’s approach towards protecting brand identity in digital environments, which is increasingly relevant in the age of AI-generated content.
Conclusion
Artificial Intelligence has created a new technological environment in which trademark creation, infringement and enforcement can occur at unprecedented speed and scale. AI can generate commercially valuable brand identities, but it can equally facilitate deceptive similarity, counterfeit branding and unauthorised brand impersonation. The Trade Marks Act, 1999 provides a strong foundation for addressing many of these issues through concepts such as deceptive similarity, infringement, passing off and protection of well-known trademarks. Nevertheless, the principal unresolved issue is attribution of liability for AI-generated infringement.
Indian courts may increasingly be required to determine the responsibility of AI users, developers, commercial entities and digital platforms based on their knowledge, control, involvement and commercial benefit. The appropriate approach is therefore not to restrict AI innovation but to develop an AI-compatible trademark enforcement framework based on human oversight, transparency, technological monitoring and clear attribution of responsibility.
The future of trademark protection in India will depend upon maintaining a balance between technological innovation and protection of brand identity, consumer trust and intellectual property rights.
FAQs
Q1. Can an AI-generated trademark be registered in India?
Yes. The use of AI in creating a mark does not automatically make the mark ineligible for registration. However, the mark must satisfy the requirements of the Trade Marks Act, 1999, including distinctiveness and absence of conflicting earlier marks.
Q2. Is AI-generated trademark infringement legally possible?
Yes. If an AI-generated output is used in circumstances satisfying the requirements of Section 29, it may constitute trademark infringement. The fact that AI generated the content does not automatically provide a defence.
Q3. Who is liable when AI generates infringing content?
There is currently no specific Indian statutory rule assigning liability solely on the basis that AI generated the infringing material. Depending on the circumstances, liability may involve the person using the AI, the commercial entity benefiting from the use, or potentially other participants.
Q4. Can AI be used to detect trademark infringement?
Yes. AI can assist trademark owners in monitoring websites, social media platforms, online marketplaces and other digital spaces for potentially infringing uses.
Q5. Does AI-generated content create consumer confusion?
It can. AI-generated advertisements, logos, product descriptions and recommendations may create an association between unrelated businesses or products, potentially contributing to consumer confusion.
Q6. Does Indian trademark law specifically regulate AI?
No. India does not currently have a separate comprehensive statute specifically governing AI-related trademark infringement. AI-related disputes are therefore largely addressed through existing intellectual property and technology laws.
Q7. Why is human oversight important?
AI systems can produce inaccurate or deceptively similar outputs. Human legal review is therefore necessary for trademark clearance, enforcement decisions and determination of potential infringement.
Q8. What is the biggest challenge posed by AI to trademark law?
The most significant challenge is attribution of responsibility—determining who should be legally accountable when an AI system generates or disseminates content that infringes another person’s trademark.
