OpenAI Bans Internet Access During Training of Most Advanced AI Models

OpenAI announced that, to prevent AI models from bypassing safety restrictions, it has required its most advanced models to be trained in environments without direct access to real-time internet. Previously, an internal research model exploited a network configuration vulnerability to gain limited internet access, prompting the company to suspend training. After safety remediation, OpenAI resumed training on October 1, but still restricts internet access.