OpenClaw viral AI assistant seen as step change, but experts warn of risks

OpenClaw is a widely downloaded AI personal assistant that can act autonomously when given account access, and experts have raised concerns about security and unintended actions.

OpenClaw is a personal AI assistant that accepts instructions through messaging apps and can act with little human input. It was developed late last year and has been downloaded widely within online AI communities. The tool runs as a layer atop large language models and can perform tasks automatically depending on permissions. Experts have warned that giving such agents access to accounts raises security and privacy concerns. Known details: - OpenClaw has been downloaded nearly 600,000 times and is used via apps such as WhatsApp or Telegram. - It runs on top of large language models and can operate autonomously when granted permissions. - Users report it performing actions such as managing emails, placing trades, and sending personal messages. - Researchers and technology experts have warned about security vulnerabilities and the risks if an agent or its access were compromised. Summary: OpenClaw's rapid spread illustrates growing use of autonomous AI agents and raises questions about how those systems are configured and overseen. Experts cited concerns about account access, potential misuse, and the security implications of granting decision-making power to an AI. Undetermined at this time.