China Issues OpenClaw Usage Guidelines, Focusing on Cybersecurity Risk Management | WPN World People News

Authorities in mainland China have recently issued their first official security guidelines regarding the AI agent tool OpenClaw, drawing widespread attention. This open-source software has gained rapid popularity in recent months, but it also comes with potential cybersecurity risks. In response, relevant regulatory bodies, following earlier risk alerts, have consolidated and released practical operational recommendations targeting general users, enterprises, cloud service providers, and technical personnel.

According to official guidance, individual users are advised to run the program on dedicated devices or within virtualized environments, keeping it isolated from daily work systems and avoiding direct installation on office computers. Users should not operate it with high-privilege accounts and should avoid processing or storing sensitive information within the environment. Keeping the software up to date is also emphasized to reduce vulnerability risks.

At the enterprise level, the importance of establishing internal governance frameworks is highlighted. This includes conducting security assessments prior to deployment, setting clear usage policies and approval processes, and strengthening system protection measures such as deploying intrusion detection mechanisms and maintaining operation logs for auditing and traceability. Regular staff training and incident response drills are also recommended to enhance overall cybersecurity awareness.

For cloud service providers offering infrastructure support, authorities require strengthened host security inspections and protective capabilities, along with improved supply chain and data protection measures to prevent potential attacks or data breaches.

As concerns over risks continue to rise, multiple institutions and companies in mainland China have begun imposing restrictions on the use of the tool, with some schools even banning it on campus. Meanwhile, some users, worried about personal data security, have started removing the software. Experts note that while such AI agents offer high levels of automation and efficiency, careful evaluation of usage scenarios is still necessary before broader adoption, particularly given that permission controls and security mechanisms are not yet fully mature.