The number of OpenClaw users has surged at the beginning of 2026. This open-source agentic artificial intelligence application, released only in November last year, has recorded more than 2 million visits in a single week.
Like other Artificial Intelligence products that carry both potential and risks, OpenClaw requires special attention from its users. The Chinese government has reportedly banned government institutions, state-owned enterprises, and major banks from installing OpenClaw on office devices. Agentic AI is considered risky, with risks ranging from cybersecurity vulnerabilities to data breaches.
Professor Ridi Ferdiana of the Department of Electrical Engineering and Information Technology (DTETI), Faculty of Engineering, Universitas Gadjah Mada ( FT UGM), explained that OpenClaw represents Agentic AI as an advanced form of artificial intelligence capable of strategic planning, complex decision-making, and autonomous task execution. OpenClaw operates by utilizing users’ internal data and external data from the internet to complete assigned tasks.
Furthermore, he noted that this Agentic AI is open-source, meaning its source code and development processes are publicly accessible, unlike closed-source systems provided by vendors.
“This is where security gaps emerge, creating the possibility of cyberattacks or data breaches, both at the individual and corporate levels,” Professor Ridi said on Sunday (Apr. 5).
According to him, being open source means OpenClaw is not only free but also transparent. Anyone can examine how it is built, enabling many developers to study and attempt to create similar systems. The information technology expert added that many users do not fully understand the information displayed in system configurations and tend to ignore prompts to update their devices.
“For those unfamiliar with the technology, permission requests on devices are often overlooked or approved without careful consideration. This is where the risk of data leakage arises, particularly for users of open-source Agentic AI such as OpenClaw,” he explained.
To address these threats, users can take several measures to minimize potential risks that could cause losses for individuals and organizations alike. Professor Ridi emphasized that users must understand the extent of their needs when using OpenClaw.
“Most of the services offered by OpenClaw are already available through third-party platforms and cloud providers, where vital data is generally more secure and protected,” he noted.
He also stressed the importance of ensuring that users’ systems, devices, and servers are secure. Once security is established, users should take extra care to read and understand permission settings and application configurations to ensure they are correctly implemented. Third, users are advised to monitor for potential data breaches at least once every two months. This is crucial, considering OpenClaw’s automation features may create scenarios beyond the control of individuals or organizations.
“We can always limit access to our private data. The key lies in extra effort, extra reading, extra updating, and extra monitoring. Security gaps can affect anyone, whether individuals or organizations; the difference lies only in the value of the data that may be exposed,” he concluded.
Author: Ika Agustine
Editor: Gusti Grehenson
Post-editor: Zabrina Kumara
Photo: Freepik