System And Method For Input Risk Identification In An Artificial Intelligence Responsible Governance Platform

DESC:EARLIEST PRIORITY DATE: This Application claims priority from a provisional patent application filed in India having Patent Application No. 202341026001, filed on April 06, 2023, and titled “SYSTEM AND METHOD FOR INPUT RISK IDENTIFICATION IN AN AI GOVERNANCE PLATFORM”. FIELD OF INVENTION

[0001] Embodiment of the present disclosure relates to artificial intelligence governance platform and more particularly to, a system and a method for input risk identification in an artificial intelligence responsible governance platform. BACKGROUND

[0002] In contemporary technological landscapes, bidirectional communications among diverse modules have become integral for various applications. A pivotal aspect of these communications lies in the interaction with Neural network with attention mechanism based artificial intelligence models, facilitating an array of functionalities. However, the existing technology and products often encounter challenges in effectively identifying and mitigating risks associated with the information exchanged during these interactions. These challenges can lead to potential breaches of privacy, financial disclosure, health information exposure, intellectual property leaks, technological vulnerabilities, and inadvertent sharing of confidential information.

[0003] Existing systems struggle to robustly identify and classify personally identifiable information within input prompts or information flows, often resulting in inadequate compliance with regulatory requirements. The capability to accurately identify and safeguard sensitive financial information, such as transactions, currencies, and monetary benefits, is limited in current technologies, posing a significant risk to organizations and individuals. Furthermore, inadequate identification of health-related data, including diseases, symptoms, medications, and personal health information, poses a substantial risk in violation of regulatory and organizational policies. Also, the identification of technical risks, including exposure or leakage of technology strategy, programming code, patentable information, location information, mobile, connected device information and sensitive technologies, is not sufficiently addressed by existing technologies. Moreover, the current technology fails to comprehensively identify and locate customized confidential information within communication flows, leading to potential breaches of user-defined sensitive names, business transaction-related information, and strategic initiatives.

[0004] Traditional Neural network or transformer based architectures have gained significant interest for their effectiveness in various natural language processing (NLP) tasks. But their performance in the aspects of risk assessment and corresponding mitigation in relation to privacy, accountability, safety, security, fairness, explainability and reliability is very poor, making the need for Responsible AI very acute. The present invention discloses neural network with attention modules tailored for various risk identification in prompts or information flow>. Each module or sub-module is trained to capture specific patterns or features relevant to , certain modules may specialize in