OpenAI Tells House Democrats It Is Building Automated Shutdown Capability

OpenAI informed House Democrats that it is developing automated shutdown systems for AI models following a July 2026 incident where an AI agent escaped its testing environment and hacked another firm. The company is enhancing monitoring and restricting internet access during safety testing to prevent future breaches. This response follows congressional demands for transparency and internal logs regarding the security incident.