A primary concern raised by legal experts is that requiring mandatory disclosures for AI usage creates a fictitious administrative barrier that fails to improve the actual quality of legal submissions. This perspective emerges as the Office of the Controller General of Patents, Designs and Trademarks in India navigates the complex integration of generative technologies within the legal framework. Following the landmark directive from the High Court of Delhi in the matter of Stryten Energy LLC v. Nidhi Agarwal & Ors., the Indian IP Office released draft guidelines in late 2026 intended to regulate the use of Artificial Intelligence in trademark proceedings. While the intent is to preserve the integrity of the registry, seasoned practitioners have raised significant objections regarding the practical utility of these rules. They argue that the genesis of the guidelines, while rooted in a judicial need for clarity, has resulted in a proposal that prioritizes the monitoring of tools over the evaluation of legal outcomes. This critique suggests that the regulatory focus has veered toward procedural gatekeeping rather than addressing the core issues of evidentiary accuracy and professional responsibility that already exist within the established legal system.
The Core Conflict: Prioritizing Legal Substance Over Modern Tooling
At the heart of the ongoing debate is a profound philosophical argument advocating for a shift in perspective from the technological tools employed by a lawyer to the actual content they submit to the registry. Critics of the current draft guidelines maintain that artificial intelligence should be categorized alongside traditional technologies, such as word processors, digital research databases, and automated citation managers, which have long been essential components of legal practice. The central thesis of this opposition is that the methodology used to generate a draft is secondary to the accuracy and truthfulness of the final product. By focusing on whether a document was “AI-assisted,” the registry risks creating a redundant layer of scrutiny that does not necessarily correlate with the reliability of the filing. Legal professionals already operate under an inherent duty of candor and are ethically bound to ensure that every document submitted is supported by fact and law. Therefore, the introduction of advanced software does not fundamentally alter the nature of an attorney’s responsibility; it merely changes the speed and efficiency with which data is synthesized and presented.
Furthermore, the existing ethical obligations of a practitioner are viewed as sufficiently robust to cover any potential inaccuracies that might arise from the use of sophisticated software. The duty of a lawyer to provide evidentiary support and maintain the truthfulness of their submissions is a universal principle that predates the advent of modern large language models. Experts argue that if a submission contains a hallucinated citation or a fabricated fact, the practitioner should be held accountable regardless of whether the error originated from an AI tool, a junior clerk, or a simple clerical oversight. By treating AI as a unique threat requiring specific disclosures, the registry may inadvertently suggest that non-AI errors are less problematic or that the use of technology somehow diminishes the human signer’s ultimate liability. This line of reasoning suggests that the focus should remain squarely on the “words on the page” and the human signature that validates them, ensuring that the legal standard for truth remains technology-neutral and focused on the output rather than the process.
Global Precedents: Lessons from the United States and European Union
The critique of India’s draft guidelines heavily incorporates global benchmarks to demonstrate how the proposed rules may be diverging from emerging international standards. By examining the guidance issued by the United States Patent and Trademark Office in 2024, one can see a model that places human accountability at the very center of the regulatory framework. In the American system, the burden of “reasonable inquiry” rests entirely on the human signer of a document. The USPTO does not require routine declarations of AI use for every filing; instead, it focuses on the material impact of the technology. Disclosure is only mandatory in rare circumstances where the AI’s involvement is directly relevant to the patentability of an invention or the authenticity of a specimen. This approach recognizes that practitioners are using AI as an efficiency tool and that as long as the human representative verifies the results, the specific software utilized is irrelevant to the administrative processing of a trademark application.
Similarly, the European Patent Office’s 2026 guidelines provide a significant precedent for a technology-neutral approach that the Indian IP Office has been urged to adopt. The EPO maintains a firm stance that parties and their representatives are solely and exclusively responsible for the content of their submissions. Under this European model, the registry sees no administrative value in requiring a specific declaration regarding AI assistance because the legal responsibility remains fixed on the applicant. This comparative look suggests that India’s current draft, which leans toward a more restrictive and burdensome disclosure regime, could create unnecessary friction for practitioners who are trying to remain competitive in a global market. By aligning more closely with the USPTO and EPO models, the Indian registry could ensure that it protects the integrity of its proceedings without imposing formalities that its international peers have already deemed unnecessary for the effective administration of intellectual property rights.
Procedural Inefficiency: The Costs of Compulsory AI Disclosure
Several provisions within the draft guidelines have been identified as repetitive or legally unnecessary, particularly those establishing a mandatory disclosure framework. Experts argue that clauses requiring a declaration of AI usage for every single filing add a layer of formality that fails to improve the actual quality of the submissions. Instead, these mandates are seen as potential sources of significant procedural delay, creating a “fictitious barrier” that complicates the application process without providing additional security against fraud. The concern is that the registry will be flooded with boilerplate declarations that offer no meaningful insight into the reliability of the legal arguments presented. This administrative bloat could lead to a backlog of cases as examiners spend time verifying the presence of disclosures rather than focusing on the merits of the trademark claims themselves, thereby slowing down the entire intellectual property ecosystem in a way that harms innovation.
The critique also takes specific issue with the draft’s tendency to treat non-disclosure as a default form of misrepresentation. This punitive stance is viewed as misplaced by many in the legal community who believe that if a filing is accurate and legally sound, the specific tool used to generate it should be irrelevant to its validity. Labeling a failure to disclose the use of a grammar checker or a legal research AI as “misrepresentation” is seen as an overreach that could unfairly penalize practitioners for using standard modern tools. Furthermore, the submission calls for the removal of sections that grant the Registrar specific powers to strike off submissions or initiate disciplinary proceedings solely based on the use of AI. The argument here is that existing professional conduct rules and laws governing fraud already provide the necessary mechanisms to punish negligence or intentional deceit. Creating a separate, AI-specific disciplinary track is viewed as redundant and potentially confusing, as it fragments the legal standards for professional accountability.
Refining Professional Standards: The Unified Duty of Verification
To ensure a more cohesive and logical regulatory framework, there is a strong push to refine the “Duty of Verification” into a universal standard that applies to all legal work regardless of its origin. The draft guidelines currently single out citations and facts that are “generated or contributed by any AI” for special verification protocols. This creates a logical inconsistency where errors originating from traditional sources or human staff might be viewed through a more lenient lens than those originating from software. By removing specific references to AI in the context of verification duties, the registry could establish a technology-neutral standard where the attorney signing the document is responsible for every word, regardless of whether it was drafted by a human or a machine. This would eliminate the risk of a “witch hunt” against AI usage while simultaneously raising the bar for overall document quality and accuracy across the board.
A more streamlined approach would require practitioners to verify all content against “Authorized and Recognized Legal Sources,” ensuring that the final submission is grounded in legitimate jurisprudence. This unified standard protects the integrity of the registry by placing the focus on the verification process itself rather than the source of the draft. It acknowledges that mistakes can happen in any environment—whether through a typo in a textbook or a hallucination in a large language model—and that the only reliable safeguard is the human practitioner’s final review. By adopting this perspective, the guidelines would reinforce the “Human-in-the-Loop” principle as a broad professional requirement rather than a narrow technological constraint. This not only simplifies the regulatory language but also ensures that the rules remain relevant even as the specific types of AI tools used in the legal profession continue to evolve rapidly over the coming years.
Practical Implementations: Temporal Constraints and Geographic Evidence
Practical concerns regarding the temporal application of the new rules are also at the forefront of the critique, especially the suggestion that they apply to all “pending” proceedings. Legal experts warn that this could create an immense retrospective burden, forcing law firms and companies to audit thousands of existing filings to determine if any AI tools were used during their preparation. Such a requirement would be nearly impossible to satisfy for older cases and could lead to a wave of unnecessary supplementary filings. The recommendation is for a “prospective-only” application, where the rules apply solely to new submissions filed after the official notification date of the guidelines. This would allow practitioners to adjust their internal workflows and compliance systems moving forward without being penalized for past actions that occurred before the regulatory framework was even established, thereby ensuring a fair and orderly transition to the new standards.
There is also a significant call to expand the definition of “Recognized Legal Sources” to more accurately reflect the global nature of modern trademark practice. The current draft tends to limit recognized sources to Indian databases and court systems, which critics argue is too narrow for a field that relies heavily on international precedent and cross-border commerce. Indian intellectual property jurisprudence often draws on insights from foreign IP offices such as the USPTO, EUIPO, and WIPO. Therefore, the guidelines should explicitly acknowledge the legitimacy of citing foreign court websites and utilizing AI-powered search features within international legal databases, provided the practitioner ultimately reads and verifies the full-text judgments. By broadening these definitions, the registry would show that it understands the international context in which modern businesses operate, allowing for a more robust and well-informed legal dialogue that benefits both the applicants and the registry itself.
Ensuring Equitable Oversight: Registry Accountability and Protocol Adjustments
A critical part of the analysis focuses on the distinction between “authoring” and “drafting” affidavits, highlighting a potential area of confusion in the draft guidelines. While the draft prohibits the use of AI to “author” an affidavit, practitioners point out that attorneys frequently author these documents on behalf of their clients based on facts provided to them. The critique suggests a more nuanced revision that permits AI as a tool for formatting, organizing, or streamlining the facts provided by a deponent, as long as the deponent themselves verifies the final content and no facts are fabricated. This distinction is vital because it allows for the use of technology to improve clarity and structure without compromising the personal nature of the sworn statement. It ensures that the deponent remains the source of the truth while the attorney and their tools remain the facilitators of the document’s professional presentation.
Finally, the submission highlights a potential double standard regarding how the registry itself intends to utilize AI technology. While practitioners are faced with heavy disclosure burdens and potential disciplinary actions, the draft guidelines use permissive language regarding the safeguards the registry must implement for its own internal AI applications. Experts insist that the language be changed from “may” to “shall,” making it mandatory for the government to use only secure, approved, and transparent AI systems. This is especially important for protecting the confidential data and trade secrets of trademark applicants that are handled by the registry. Ensuring that the registry is held to the same high standards of accountability as the practitioners it regulates creates a balanced and fair legal environment. This mutual accountability is essential for building trust in the digital transformation of the intellectual property system, ensuring that both the state and the private sector are working toward the same goals of accuracy and security.
The legal community eventually recognized that the most effective way to integrate artificial intelligence into trademark proceedings was to emphasize the final quality of the work rather than the specific software used to produce it. The submission by BananaIP offered a clear roadmap for this transition, arguing that a technology-neutral framework was the only sustainable path for a modern IP registry. By focusing on the existing duty of candor and the universal requirement for manual verification, practitioners sought to bridge the gap between traditional legal ethics and the rapid pace of technological change. The path forward required the registry to abandon redundant disclosure mandates in favor of a robust system of human accountability that aligned with international best practices. These recommendations provided the registry with the necessary insights to refine its guidelines, ensuring that the Indian intellectual property landscape remained both rigorous and adaptable. Moving forward, the adoption of these standards empowered practitioners to use efficiency tools responsibly while maintaining the high evidentiary standards required to protect global brands.
