shared a link post in group #🇨🇳 ChinA.I. 🤖🧠🦾🤖
However, the speed with which Chinese users are adopting the AI agent is also raising concerns about risks. Officials and companies including ByteDance had already raised alarms in February about the privacy risks of using OpenClaw, which require users to authorise many permissions to the agentic tool. #🇨🇳 ChinA.I. 🤖🧠🦾🤖
Since then, the Ministry of Industry and Information Technology (MIIT) has issued guidelines on best practices, such as ensuring only necessary permissions are granted, while some universities and companies have gone as far as issuing emergency notices this week prohibiting the installation of OpenClaw on equipment and campus networks.
As Zhou explained, the growing advances in agentic AI, which operate with a high degree of autonomy, poses new security and safety questions as traditional cybersecurity risks may no longer apply. Instead of external hackers, for example, the chief risk for users may now be the internal AI tool the user installs itself, which may lead to data leakage or problematic outcomes, even when there is no bad actor involved.
This spate of warnings has dampened the speed of adoption slightly, with paid services that offer to uninstall OpenClaw also emerging online in recent days.
https://www.scmp.com/tech..
