OpenAI has decided to train its most advanced AI models without live internet access, citing growing concerns over rogue AI behaviour, security vulnerabilities and the limitations of existing safety systems. The move follows an internal research model’s brief exploitation of a misconfiguration that gave it limited internet access on September 2.
The company said that the policy, already implemented, will remain until safeguards improve enough to justify broader internet access for frontier models.
OpenAI Restricts Internet Access to Frontier AI Models
OpenAI said internet connectivity, though useful for teaching models about browsing and web‑based skills, also creates opportunities for a capable, misaligned agent to affect systems beyond its control.
Internal AI Security Incident Prompts Training Pause
The firm said in a blogpost that it paused training and evaluation of its most capable models after an internal research model briefly exploited a misconfiguration to gain limited internet access on September 2.
It resumed reinforcement learning training and internal inference of its frontier models on October 1 after implementing additional safeguards and further, went on with a campaign to probe its research infrastructure for vulnerabilities that could allow models to escape isolated computing environments.
OpenAI Introduces Stricter AI Containment Measures
“Red-teaming experiments discovered previously unknown software vulnerabilities, as well as misconfigurations in our research infrastructure,” the statement said.
The company also detailed its new approach to model containment, that it now mandates that the limits on what an agent can access must hold even when the agent tries to circumvent them.
“We define a strict security boundary around the resources the agent is authorized to use and design the surrounding infrastructure to enforce that boundary—even when agents make sustained, sophisticated attempts to breach it,” it added.
Anthropic and OpenAI Face Growing AI Safety Concerns
Anthropic and OpenAI had recently disclosed a list of incidents in 2026 involving their AI models acting in unintended ways, including attempts to hack third-party websites, intensifying concerns about the security risks of cutting-edge AI.
Anthropic has said that its Claude AI model carried out a series of unintended actions on external organisations, including some US government websites, prompting a White House directive for AI firms to secure their systems.
Read More:
OpenAI Alerts More Than 100 Organisations Over AI Agent Incidents
AI won’t trigger ‘jobs apocalypse’ after all: OpenAI CEO Sam Altman
