MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611059364 A) filed by Anuj Tyagi; and Swati Tyagi on May 10, 2026, for System And Method For Securing A Large Language Model Application.
Inventors include Anuj Tyagi; and Swati Tyagi.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: ABSTRACT SYSTEM AND METHOD FOR SECURING A LARGE LANGUAGE MODEL APPLICATION Disclosed are a system (100) and method (200) for securing a Large Language Model (LLM) application are disclosed. The system (100) includes a user device (102) and a server (104) having processing circuitry (124). The processing circuitry (124) receives a natural-language input prompt from a client application through the user device (102) and analyzes the input prompt using a plurality of prompt scanners (130A-N) to detect threat vectors including prompt injection, sensitive information disclosure, toxicity, invisible text manipulation, data poisoning, supply chain vulnerabilities, and resource abuse. Structured scan results are aggregated using a selectable fail-fast mode or fail-safe mode to determine whether the input prompt is forwarded to a target LLM. An output response from the target LLM is analyzed by a plurality of output scanners (132A-N) for detecting contextual irrelevance, factual inconsistency, data leakage, or malicious content. The processing circuitry (124) maps scanners to an OWASP LLM Top 10 taxonomy, computes a compliance score, and selectively transmits either a validated response or a structured security response with remediation guidance. Figure 1 will be the reference
Disclaimer: Curated by HT Syndication.
