TScientific Achieves GS Certification Grade 1 for AI-Based Personal Data Detection Solution

TScientific announced on the 14th that its AI technology-based unstructured personal information detection solution, ‘P3R-Privacy Finder AI,’ has obtained the GS (Good Software) Certification Grade 1.

The GS Certification is a software quality certification system that officially certifies products meeting certain quality requirements such as functional suitability, performance efficiency, usability, and reliability based on international standards in testing and evaluation conducted by the Software Testing and Certification (KTL, Korea Testing Laboratory) division.

Software that obtains GS Certification is eligible to apply for selection as an excellent procurement item by the Public Procurement Service and designated as a priority purchase product for public institutions, among other benefits.

This product applied large language model (LLM) technology specialized in large-scale language analysis, which was researched and developed over three years, and developed a language model specialized for Korean. It was also developed as an open application programming interface (API) that can be utilized in other solutions or systems, making it convenient to apply to solutions and systems requiring personal information protection.

In particular, the company explained that messengers, chatbots, and business systems specialized in detecting personal information in conversational sentences and documents are expected to be linked with various services in the future to prevent personal information leakage incidents.

Lee Jong-gyu, head of the TScientific Research Institute, said, “After much effort to find a suitable AI model and optimize it, we succeeded in productizing it. We will do our best to ensure that the related model can be used in various services requiring personal information protection.”

He added, “We will not stop at this commercialization success but will continue research and development (R&D) related to identifying true and false positives to improve the accuracy of personal information protection, enhance usability, and support more convenient use.”