AI security company Raon Secure announced that its AI security solution, OnShield, has received the AI Product and Service Certification as of October 14.
The AI Product and Service Certification system, established by the Korea Artificial Intelligence and Software Industry Association (KOSA), verifies whether AI technology is actually applied to products or services in accordance with the AI Basic Act. The certification is issued after a technical review of the AI's learning and inference characteristics and its functional contributions.
As the adoption of generative AI increases, new security threats have emerged that existing security systems struggle to address. Examples include employees inputting sensitive company information or personal data into generative AI during work, and attackers using malicious prompts to bypass security policies set for AI.
OnShield is an AI safeguard solution designed to counter these threats. It analyzes input and output data exchanged between users and generative AI or AI agents in real time, allowing or blocking data based on pre-set policies. This helps manage the risks of data leaks and AI misuse that may arise when companies utilize generative AI.
Key features include detection and blocking of malicious prompt attacks such as prompt injection and jailbreaking, as well as automatic identification, anonymization, and masking of personal information included in AI input data. It filters harmful content based on a Korean-language specialized model and controls the input of sensitive information and inappropriate AI usage according to security policies set by organizations.
With the acquisition of this certification, Raon Secure plans to intensify its efforts to penetrate the public and financial markets. The certification is expected to serve as a basis for public institutions and financial organizations to assess the application of AI technology in OnShield as they consider adopting AI security solutions.
Lee Jeong-a, CEO of Raon Secure, stated, “As we transition from utilizing generative AI to having AI agents perform actual tasks, new security threats such as sensitive information leaks, malicious prompts, and policy circumvention are also increasing. OnShield is a solution that controls these threats, enabling companies and institutions to use AI more safely. With this certification, we aim to expand our business opportunities, focusing on the public and financial sectors.”
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.