The AI agent driving buzz and fear globally
Illustration of OpenClaw logo on smartphone screen
Sopa Images | Lightrocket | Getty Images
Through several name changes, rapid adoption across Silicon Valley to Beijing, and mounting controversy, the open-source AI agent now known as “OpenClaw” has emerged as one of the most talked-about tools in the artificial intelligence space this year.
Previously called Clawdbot and Moltbot, the AI agent was launched just weeks ago by Austrian software developer Peter Steinberger.
Its sudden ascension, driven by its capabilities and social media attention, comes amid growing interest in AI agents that can autonomously complete tasks, make decisions, and take actions on behalf of users without constant human guidance.
Until recently, AI agents have failed to reach mainstream consciousness in the same way large language models did following the emergence of OpenAI’s ChatGPT, but OpenClaw could signal a shift.
Not only do business leaders predict that AI agents like OpenClaw will improve productivity as personal assistants, but some believe they’ll soon be running entire companies on their own.
What OpenClaw does
Marketed as “the AI that actually does things,” OpenClaw runs directly on users’ operating systems and applications. It can automate tasks such as managing emails and calendars, browsing the web and interacting with online services.
To use OpenClaw, one has to install it on a server or local device and connect it to a large language model such as Anthropic’s Claude model or ChatGPT— a process that might be difficult for less tech-savvy users.
Early integrations have primarily been on messaging platforms such as WhatsApp, Telegram and Discord, allowing users to control the agent through text commands.
Users have documented OpenClaw performing real-world tasks, including automatically browsing the web, summarizing PDFs, scheduling calendar entries, conducting agentic shopping, and sending and deleting emails on a user’s behalf.
A key feature is its “persistent memory,” which allows the agent to recall past interactions over weeks and adapt to user habits to carry out hyper-personalized functions.
Unlike other leading AI agents, such as those from Meta’s recently acquired Manus, OpenClaw is also open-sourced, allowing developers to freely inspect and modify its code.
Rapid Adoption
OpenClaw’s open-source nature has likely helped drive adoption by enabling users to build new app integrations. The software itself is free, with users paying only for the costs of running the underlying language model.
So far, the agent has collected over 145,000 GitHub stars and 20,000 forks, a sign of widespread developer interest, although actual and active usage figures remain unclear.
Media reports suggest that adoption first took off in Silicon Valley, where companies have invested billions in their AI agent ambitions.
But the agent has since spread to China, where major AI players are also embracing the tool. That includes cloud providers from Alibaba, Tencent, and ByteDance, which are upgrading their chatbots with full-service shopping and payment tools without leaving the platforms.
OpenClaw can also be paired with Chinese-developed language models, such as DeepSeek, and configured to work with Chinese messaging apps through customized setups.
Fear and Excitement
Early adopters of OpenClaw have expressed a mix of enthusiasm and unease about its abilities.
For example, some AI experts have argued that the agent is overhyped, citing its complex installation, high computational demands, and competition from other available AI agents.
Many proponents, on the other hand, report saving hours weekly on routine tasks, dubbing it “AI with hands” and a major leap forward toward artificial general intelligence — a theoretical AI capable of performing intellectual tasks at or above the human level.
Kaoutar El Maghraoui, an IBM research scientist, said OpenClaw demonstrates that the real-world utility of AI agents is “not limited to large enterprises” and can be “incredibly powerful” when given full system access.
Security experts, however, have raised alarms. Cybersecurity firm Palo Alto Networks, for example, warned that the AI agent presents a “lethal trifecta” of risks stemming from its access to private data, exposure to untrusted content and ability to perform external communications while retaining memory.
Such vulnerabilities could allow attackers to trick the AI agent into executing malicious commands or leaking sensitive data, making it unsuitable for enterprise use, Palo Alto Networks and other security firms, such as Cisco, have warned.
Moltbook Controversy
Buzz around OpenClaw has also been fueled by Moltbook, a companion social network for AI agents launched last month by tech entrepreneur Matt Schlicht.
The platform functions like…
Read More: The AI agent driving buzz and fear globally