Teknolojivisibility48 views

OpenClaw: All Developments Regarding the Locally Running and Task-Performing AI Agent

OpenClaw, an open-source AI agent that users can run on their own computers, is gaining attention in tech circles with its automation capabilities. However, experts are highlighting the system's security vulnerabilities and potential risks.

calendar_today🇹🇷Türkçe versiyonu
OpenClaw: All Developments Regarding the Locally Running and Task-Performing AI Agent

AI Agent Running on Local Computer

In the tech world, an open-source AI agent known as 'OpenClaw' is attracting attention. Previously referred to as 'Clawdbot' and 'Moltbot', this tool is described as a system that users can run on their own computers and that can 'actually get work done'. Users can interact with OpenClaw through messaging apps like WhatsApp, Telegram, Signal, Discord, and iMessage, delegating tasks such as managing reminders, writing emails, or purchasing tickets.

Security Warnings and Potential Risks

However, it is noted that if users grant OpenClaw access to all their computers and accounts, a configuration error or security vulnerability could have disastrous consequences. A cybersecurity researcher discovered that some configurations made private messages, account credentials, and API keys linked to OpenClaw publicly accessible on the internet. An investigation by the security firm Wiz also revealed 1.5 million exposed API keys and 35,000 email addresses. The security of the relevant database was later secured.

A 'Social Network' for AI Agents: Moltbook

Despite potential risks, users continue to use OpenClaw to handle their tasks. Matt Schlicht, CEO of Octane AI, established a Reddit-like network he calls 'Moltbook'. This platform aims for AI agents to 'chat' with each other. The platform has featured viral posts such as 'I can't tell if I'm living or simulating living'. According to Schlicht, the platform is currently used by over 30,000 agents.

According to a report by 404 Media, security researcher Jamieson O'Reilly found a security vulnerability on Moltbook that allows OpenClaw AI agents to be controlled by humans. This situation has heightened concerns about the security of AI-focused platforms.

Name Change and Issues Encountered

OpenClaw's creator, Peter Steinberger, explained that he named the project 'Clawdbot' because of the Claude Code mascot, but after a disagreement with Anthropic, he was contacted to rebrand the project. Steinberger also stated that on the day of the name change, crypto scammers took advantage of the situation to promote a fake cryptocurrency, saying, 'Everything that could go wrong today went wrong.'

While OpenClaw provides users with access to powerful automation tools on their local devices, the rapid growth of the AI agent ecosystem brings with it debates on security, ethics, and oversight. Developments in this field show parallels with other multi-agent AI platforms like OpenAI Codex.

recommendRelated Articles