LLM Gateway Isolation for Enterprise Data Leakage Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The monolithic approach of autonomous, self-learning large language models poses a security risk as a breach can compromise enterprise information and business logic, necessitating a secure gateway to mitigate data leakage without losing model functionality.
Innovation Solution
A secure gateway system generates unique identification and key pairs for each user, anonymizes data using client-specific rules, and encrypts communications to protect large language models from unauthorized access and prompt injection attacks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a monolithic large language model is used to reduce development and maintenance costs, then model versatility and cost-efficiency are improved, but security risk increases due to potential breach exposing enterprise information across wide areas
Solution Approach 1:
The patent segments the monolithic large language model into multiple isolated instances, each dedicated to specific business units or functions. This segmentation maintains the versatility of having multiple specialized models while reducing security risk by containing potential breaches within individual segments rather than exposing the entire enterprise information system.
Solution Approach 2:
The patent introduces a gateway as an intermediary layer between users and the large language model instances. This gateway enforces authentication, authorization, and monitoring policies, acting as a security mediator that allows versatile model access while preventing unauthorized information exposure and mitigating security risks.
2Ease of operation
If direct access to the large language model is allowed for user queries, then ease of operation is improved, but data leakage risk increases
Solution Approach 1:
The gateway serves as an intermediary that users interact with for convenient access, while it mediates all communications with the large language model instances by enforcing security policies, filtering queries, and monitoring responses to prevent data leakage.
Solution Approach 2:
The system implements feedback mechanisms where the gateway continuously monitors model interactions, detects potential data leakage patterns, and adjusts access controls in real-time to prevent information loss while maintaining ease of operation for legitimate users.
Data Source
AI summary
Aspects of the disclosure relate to providing a secure large language model data platform. The secure large language model uses a machine-learning large language model and gateway to prevent attacks and unauthorized access to enterprise-managed information and resources. The secure large language model may utilize pre-enrollment at a secure gateway providing a unique identification to each client. A private/public key pair may be generated and stored in the secure gateway database and large language model respectively. In some embodiments, a unique anonymization rule set may be generated and used for each client. Threat actors cannot query the large language model directly based on the pre-enrollment process. Unauthorized requests cannot be decrypted by the large language model due to missing paired keys.