[Asia Economy Reporter Oh Hyung-gil] Starting from the second half of this year, a monitoring system will be introduced that collects online financial advertisements using big data to analyze their illegality and block illegal advertisements.
On the 7th, the Financial Supervisory Service (FSS) announced that as part of the digital transformation of financial supervision, it will introduce artificial intelligence (AI) and big data to establish and operate five systems, including AI private fund review, identification of incomplete sales in insurance telemarketing (TM), and monitoring of illegal online financial advertisements.
Since May last year, the FSS has applied AI technology, specifically speech-to-text conversion, to recorded files collected during inspections of debt collectors, identifying illegal debt collection practices such as verbal abuse or repeated collection attempts.
Additionally, last month, using speech conversion technology, the FSS analyzed TM recordings during insurance sales inspections to detect false information regarding mandatory disclosures to insurance policyholders and determine cases of incomplete sales.
Furthermore, by utilizing AI, the FSS analyzes complaints submitted by customers to automatically recommend complaint types or similar complaints to staff, and employs machine reading technology for AI to read private fund reports, assess appropriateness by key items, and support review tasks.
The FSS plans to establish a system by July that collects big data related to illegal financial advertisements scattered across blogs and social networking services (SNS) online and analyzes them based on rule-based methods.
This method analyzes the presence or absence of major keywords related to six types: unregistered lending, bank account trading, loan manipulation, cashing out small payments via mobile phones, cashing out credit card payments, and trading of personal credit information.
The FSS judges that the recognition rate and accuracy of AI and big data technologies are quite favorable and expects the contribution to work to gradually improve.
Currently, the speech recognition rate for recorded files is 89.5% for debt collection and 93.7% for insurance TM incomplete sales, while the machine reading accuracy used to support private fund review tasks recorded about 94.5%.
An FSS official stated, "This year, we will build a big data utilization base for the complaint consultation system and advance it into a system that can comprehensively analyze complaint trends as well as consultations. We will strengthen financial supervisory capabilities in preparation for the digital transformation era and respond more quickly to the voices of financial consumers."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
