Session on Interface Between AI and IP: The Trade Secret Perspective
Event Date and Year: 15th Jan 2026
Event Brief Description:
The School of Law recently hosted a specialized session within its Certificate Course on the Interface between AI and IP, featuring Dr. Debmita Mondal, Assistant Professor at HNLU, Raipur. An alumna of NALSAR and a PhD recipient under the guidance of Prof. V.C. Vivekanandan, Dr. Mondal brought her extensive expertise in Intellectual Property and Information Technology Law to examine the burgeoning role of Trade Secrets in the AI ecosystem.
The session underscored why trade secrecy has become the preferred vehicle for AI protection, primarily due to the limitations of patent disclosure and the rapid nature of innovation cycles. Dr. Mondal expertly navigated the "Trade Secret Perspective," detailing the tripartite requirements for protection: secrecy, commercial value, and reasonable security measures. By analysing the Indian position—which relies on contract law and equity in the absence of a standalone statute—the lecture provided a critical look at how proprietary datasets, model architectures, and weight parameters are shielded as confidential information. The session bridged the gap between theoretical common law principles and modern technological challenges, such as model extraction attacks and prompt injections.
Event Detailed Description:
The session led by Dr. Debmita Mondal provided an advanced exploration of the AI-Trade Secret Interface, a domain where confidentiality often triumphs over the public disclosure requirements of patent law. Dr. Mondal initiated the discourse by highlighting that many AI components—specifically algorithms, curated datasets, and learned weight parameters—frequently fail the stringent "non-obviousness" or "technical effect" standards required for patentability. Consequently, trade secrecy offers a "perpetual" alternative for maintaining competitive advantage.
The Indian Jurisprudential Landscape Dr. Mondal elucidated that in India, trade secret protection is rooted in Equity and Common Law rather than a dedicated statute. She analyzed the landmark case of John Richard Brady v. Chemical Process Equipments Pvt. Ltd., wherein the Delhi High Court recognized the "spring-board" doctrine, prohibiting defendants from utilizing confidential technical information to gain an unfair competitive start.
Categorization of AI Trade Secrets The lecture categorized AI-related trade secrets into three distinct tiers:
Emerging Legal and Technical Challenges A significant portion of the lecture focused on the vulnerability of AI systems to "misappropriation" via technical means. Dr. Mondal discussed Model Extraction Attacks, where an adversary approximates a surrogate model without direct source code access, and Inference Attacks, such as those alleged in Open Evidence v. Pathway Medical. The latter case raises a pivotal legal question: whether "reasonable measures" to protect a secret truly exist if a system can be compromised via a series of "prompt injections."
Litigation and Accountability Dr. Mondal examined the high-stakes world of employee mobility through cases like Waymo LLC v. Uber and Huawei v. CNEX Labs. These cases illustrate the fine line between "acquired skill and knowledge" (which an employee is free to use) and "trade secret misappropriation" (which is actionable). Furthermore, she cited T2 Modus LLC v. Williams-Arowolo to remind practitioners that courts will not accept conclusory terms like "AI" or "machine learning" as a description of a trade secret; specific technical parameters must be identified to sustain a claim.
Department Name: School of Law, Galgotias University
Event Outcome: