Korea Expert (CEO Yoo In-ji) announced last month that it had successfully completed the Korea Credit Information Service's "Financial AI Data Library System" construction project.
The “Financial AI Data Library” is a data infrastructure that allows safe reuse of pseudonymous information* (combined data) and is one of the major projects being promoted by financial authorities.
Through the revision of the Credit Information Act in 2020, pseudonymous information can be used for tasks such as AI learning, innovative service development, and advancement of credit evaluation and risk management. However, all data before and after combination must be destroyed immediately, and data use organizations are allowed to use the data only for the purposes stated when applying for data combination.
For example, even if a data specialized agency (Korea Credit Information Service, Financial Security Institute, etc.) wants to combine data of A+B, A+C, and A+D, it cannot store A, so it must receive A from the data holding agency. It must be transmitted every time, and financial companies and fintech companies cannot use the data combined for the purpose of 'advancement of credit rating model' for 'loan screening strategy', so they have to reapply for data combination every time. There were great inconveniences in building and using large, high-quality data sets.
▲ Comparison of current data combination and “Financial AI Data Library” (Comparison of Korea Credit Information Service)
Through this project, we will build a “Financial AI Data Library,” a platform that can store and utilize data before and after combination without destruction, ▶ safely reusing pseudonymous information, ▶ saving time and cost for data combination, and ▶ providing data to third parties. Provides the ability to view and reuse the data list held by .
After carrying out the Korea Credit Information Service's “Insurance Fraud Prevention Work Integration and System Efficiency” project in 2020, Korea Expert has experienced various large and small information integration, inquiry, and system connection tasks to expand the data infrastructure of the Korea Credit Information Service.
Baek Seok-hyun, director of Korea Expert, who was in charge of the PL for this project, said, “Combining data between general companies holding financial data (performed by a data specialized agency), configuring infrastructure for data combination, storage, and reuse, developing web services, sending and receiving large files “We performed work such as developing API modules for AI,” and added, “We expect that various new AI services and systems will be developed through the financial AI test bed, and that we will be able to contribute to innovation and revitalization of the AI ecosystem in the financial sector.” said.
* Pseudonymous information: Information processed so that a specific individual cannot be identified. Pseudonymous information can be used without the consent of the information subject for statistical purposes, research, and public record preservation.
Comments