Table of Contents
GenAI, or generative AI, is a fast-growing field that uses AI to create new content, such as text, images, audio, and video. However, GenAI also poses significant security risks, as it involves processing large amounts of sensitive data through APIs and large language models (LLMs).
To help organizations protect their GenAI and LLMs from cyberattacks, Akto, a leading provider of API security solutions, has launched a new platform for GenAI security testing. The platform is the world’s first proactive solution for GenAI security, according to the company.
Why GenAI security is important
GenAI is becoming more popular and widely used across various domains and applications. According to a recent Primavera report, an average organization uses 10 GenAI models, most of which are LLMs that receive data indirectly via APIs.
However, this also means that there is a lot of sensitive and personally identifiable information (PII) flowing through the LLM APIs, which can be exploited by malicious actors. LLMs can be attacked in various ways, such as denial of service (DoS), prompt injection, and overreliance on LLM outputs.
For instance, last year, OpenAI’s AI tool suffered a major outage due to a vulnerability in an open-source library, which exposed some customers’ payment information. Moreover, a study by Menlo, a browser security leader, revealed that 55% of GenAI inputs contained PII and that organizational security policies were not keeping up with the rapid evolution of GenAI.
The study also warned that the 80% increase in attempted file uploads to GenAI websites posed a new and serious cybersecurity threat. It advised organizations to update their application list or risk exposing their data and systems to GenAI sites that employees were using.
How Akto’s platform works
Akto’s new platform for GenAI security testing aims to address these threats by using state-of-the-art algorithms and over 60 built-in test cases, developed by Akto’s team of experts in GenAI security. The platform covers a wide range of GenAI vulnerabilities and helps organizations enhance the security of their GenAI and LLMs.
The platform works by identifying and sanitizing the output of LLMs, which can contain malicious code or other types of cyberattacks that can compromise the system. The platform also tests the GenAI application for the tendency to generate irrelevant or false reports, which can affect the quality and reliability of the GenAI output.
“The tests try to exploit LLM vulnerabilities through different encoding methods, separators, and markers,” said Ankush Jain, Akto’s chief technology officer. “This specially detects weak security practices where developers encode the input or put special markers around the input.”
Akto’s demo at Austin API Summit 2024
To demonstrate the capabilities of the new platform, Akto’s founder and CEO Ankita will be presenting at the prestigious Austin API Summit 2024. The demo will showcase real-world security challenges and how Akto’s platform can provide a robust defense against them.
The Austin API Summit 2024 is a premier event for API professionals and enthusiasts, where they can learn, network, and share best practices and innovations in the API industry. The summit will feature keynote speakers, workshops, panel discussions, and exhibitors from leading API companies and organizations.
Akto’s vision for GenAI security
As GenAI adoption continues to rise, Akto has positioned itself at the forefront of proactive security testing for GenAI applications. The launch of the new platform reinforces Akto’s commitment to enabling a safer environment for organizations to deploy GenAI.
Akto’s vision is to empower organizations to leverage the benefits of GenAI without compromising their data and systems. By providing a comprehensive and easy-to-use platform for GenAI security testing, Akto hopes to help organizations achieve their GenAI goals with confidence and peace of mind.
