Major South Korean tech companies including Kakao are restricting developers' use of "OpenClaw" as part of enhanced security measures. The move comes amid concerns that sensitive corporate secrets and personal information could be leaked externally when OpenClaw directly controls users' computers.
According to IT industry sources on the 8th, Kakao recently announced restrictions on OpenClaw use within its internal network and work devices to protect information assets. Naver and Karrot have also decided to block access to OpenClaw and MoltBot within their companies. This marks the first official ban on a specific AI tool in approximately one year, following restrictions imposed by public institutions and corporations on Chinese AI model "DeepSeek" in early last year.
OpenClaw is an open-source AI agent based on large language models (LLM) that automatically controls users' PCs. It has gained popularity among developers for its ability to automate repetitive tasks and collect and organize information across multiple websites. However, integration with internal corporate systems poses risks of confidential information leaking externally. Some security analyses have identified cases where API keys were stored unencrypted during OpenClaw-based agent implementation.
These concerns appear to be driving companies to proactively block internal network access or issue usage advisories.
Warnings are emerging overseas as well. China's Ministry of Industry and Information Technology (MIIT) has demanded strong identity authentication and access control systems, warning that improperly configured OpenClaw could become a data breach channel. Microsoft has also officially raised concerns about its use in enterprise environments.