
Security vulnerabilities in APIs (Application Programming Interfaces) have been exposed in recent personal information infringement (hacking) incidents involving the fashion platform Musinsa, the beauty and plastic surgery information platform Gangnam Unni, and the fan platform Weverse. As hacking cases through APIs—used for connections between platforms or service partnerships—are being reported one after another, there are calls for strengthened security measures tailored to these threats.
According to the information and communications industry and the Personal Information Protection Commission on the 28th, recent hacking incidents involving Musinsa (29CM), Gangnam Unni (Healing Paper), and Weverse share a common point: personal information was leaked through APIs. In the case of the Gangnam Unni hacking incident that occurred on the 4th, it was determined that the hacker exploited vulnerabilities in an API used to query web service consultation records. For Musinsa, it was found that the attack targeted the member order information API. Similarly, for Weverse, authorities are investigating reports that a hacker accessed payment-related APIs to leak personal information.
APIs, which are primarily used to connect platforms or services, serve as channels through which substantial personal information flows. According to traffic data aggregated by the traffic management service Cloudflare, 57% of total internet service traffic is exchanged via APIs. Consequently, in recent hacking incidents, it was determined that hackers requested information like normal users to collect the personal information of numerous users from the APIs.
The Personal Information Protection Commission points out that inadequate API management can make systems a primary target for hacking. Hackers mimic legitimate requests for information from APIs and then exploit security vulnerabilities such as personal information query permissions and call frequency limits. The Commission explained that merely checking login credentials without properly verifying personal information query permissions, or including unnecessary personal information in response data with a single information request, can lead to large-scale personal information leaks.
In particular, when processing personal information via APIs, information not verified by website users may also be exchanged. It is pointed out that even if a customer purchases a product by entering only their name on a webpage, personal information such as email addresses, phone numbers, and addresses can move together due to the nature of connected data. Other issues have emerged, including: the potential for the scale of personal information leaks to increase due to automated attacks through programs such as AI (Artificial Intelligence); security vulnerabilities arising when unused APIs are not deleted or managed; and negligence in permission management due to personnel changes such as promotions or resignations.
To manage these vulnerabilities, the Personal Information Protection Commission recommended that businesses utilizing APIs: adhere to the principle of personal information minimization from the design stage; grant and manage API permissions in accordance with the principle of least privilege; and identify, update, and inspect lists of operating APIs. Additionally, it instructed businesses that receive APIs to operate services to verify whether response data contains personal information unrelated to the service and to immediately remove any unnecessary personal information.
Lim Jong-in, a professor at the Graduate School of Information Security at Korea University, advised, "As we transition into the era of agentic AI (Artificial Intelligence), information movement utilizing APIs is bound to increase, and the risk of security incidents is also rising." He added, "Companies handling personal information must set appropriate guardrails—guidelines that prevent inappropriate actions—in the process of information movement via APIs, and actively consider adopting the latest technologies, such as abnormal behavior detection systems utilizing AI."