eSentire Labs has recently announced the release of their eSentire LLM Gateway, a framework that provides security teams with visibility and control over the use of ChatGPT and other Large Language Models (LLMs) within their environment. To improve security measures, eSentire made the LLM Gateway available on GitHub, making it an open-source implementation.
Originally developed for internal use, the LLM Gateway was created as a means for eSentire to govern and monitor the usage of ChatGPT, Google Bard, and other LLMs. With the support of this framework, security operations centers and threat response teams can now access telemetry data, which allows them to identify and counter advanced LLM-based attacks. Alexander Feick, the Vice-President of eSentire Labs, stated that these attacks surpass traditional methods such as phishing and impersonation attacks.
The open-source nature of this project provides IT administrators with the freedom to integrate and apply their own security controls, including corporate policies, usage rules, and prompts. This allows organizations to tailor the framework to their specific needs, giving them a flexible and customizable approach to LLM security.
The primary function of the LLM Gateway is to establish a protective layer between corporate data and AI applications. Through this layer, all interactions with LLMs are logged, enhancing monitoring capabilities and security controls. By storing these logs, organizations can analyze and review the usage of LLMs, ensuring that they are being utilized safely and in accordance with best practices. Additionally, the framework offers plug-ins that provide IT administrators with recommendations on how to visualize and track LLM usage effectively.
Recognizing the growing prominence of LLM-based attacks, eSentire has plans to introduce a complete managed detection and response service for LLM. This service will provide organizations with a comprehensive solution to detect and mitigate LLM-related security threats. By offering managed services, eSentire aims to alleviate the burden faced by security teams and provide them with the necessary resources to defend against LLM attacks effectively.
In the ever-evolving landscape of cybersecurity, staying up-to-date with the latest threats and vulnerabilities is crucial. To address this need, eSentire offers a newsletter subscription service, delivering the latest information on cybersecurity threats, newly-discovered vulnerabilities, data breach incidents, and emerging trends. Subscribers can have these updates conveniently sent to their email inbox daily or weekly, ensuring that they are constantly informed and prepared to tackle the ever-changing landscape of cybersecurity.
The release of the eSentire LLM Gateway marks a significant milestone in enhancing the security of ChatGPT and other LLMs. With its open-source implementation and customizable nature, organizations have greater control over their LLM usage, allowing them to safeguard their data and detect sophisticated LLM-based attacks effectively. eSentire’s commitment to providing managed detection and response services further demonstrates their dedication to supporting organizations in the face of rapidly evolving cybersecurity threats.