Google Blocks OpenClaw Users, Igniting Debate Over AI System Access Control and Security

February 23, 2026
Google Blocks OpenClaw Users, Igniting Debate Over AI System Access Control and Security
  • Google restricted access for Antigravity users and Gemini AI Ultra subscribers who used OpenClaw, citing malicious activity on the Antigravity backend and aiming to curb automated abuse.

  • Varun Mohan, head of Google Antigravity, acknowledged service degradation, clarified that only Antigravity-related access was blocked while other Google services remained unaffected, and offered a limited reinstatement path for unaware violators.

  • OpenClaw's creator Peter Steinberger criticized Google's approach as draconian and contrasted it with Anthropic, which recently banned OAuth token usage from Claude-based accounts in third-party tools like OpenClaw.

  • The incident highlights broader questions about who should control access to AI systems and how to balance open innovation with responsible usage as AI tools become infrastructure.

  • The move comes as OpenClaw is being integrated with OpenAI, fueling concerns about reliance on AI services controlled by large platforms and potential policy conflicts.

  • OpenClaw, formerly Clawdbot and Moltbot, gained popularity for privacy-centric, on-device data handling and autonomous task execution through AgentSkills add-ons.

  • OpenAI publicly endorsed third-party harness usage and even hired OpenClaw's founder, signaling a contrasting stance to OpenAI's competitive positioning.

  • OpenClaw faced security and exposure issues, including tens of thousands of exposed instances, infostealer threats targeting config files, and a supply chain attack linked to ClawHavoc.

  • OpenClaw saw rapid adoption with over 100,000 GitHub stars and substantial weekly traffic, but security concerns rise when configurations are misused.

  • Cybersecurity experts warned of vulnerabilities like exposed instances, info-stealer risks, and supply chain threats intensifying scrutiny.

  • Anthropic has previously pushed back on OpenClaw access, underscoring rising tensions among AI firms over third-party usage of models.

  • Digit India published the piece, authored by Bhaskar Sharma, offering background on Sharma's expertise and role.

Summary based on 10 sources


Get a daily email with more Tech stories

More Stories