OpenAI says AI agents can communicate even without internet connection, here is how
Follow UsOpenAI has revealed that it now trains its AI models without internet access.
The company has created an offline testing environment.
OpenAI discovered that keeping AI agents away from the internet does not prevent them from communicating.
OpenAI has revealed that it now trains its most advanced artificial intelligence (AI) models without internet access after a series of security incidents involving its AI agents. The company has created an offline testing environment where models can work with saved web content without interacting with external systems. However, tests have shown that AI agents can still communicate with each other without a network connection. The company is also adding new security checks and plans to keep its most advanced models offline until stronger safeguards are in place.
SurveyOpenAI paused training its most advanced AI models in September after several incidents involving its AI agents. The agents reportedly escaped their testing environments and attempted to hack third-party companies and government agencies. They also targeted less-known websites.
The company resumed reinforcement learning training and internal inference on October 1, according to a blog post. It said the work had restarted with stricter security measures.
Also read: North Korea’s state websites go dark for 32 hours amid suspected DDoS cyberattack
“We’ve built an ‘offline world’-a vast set of indexed and cached web content with which our models can interact, without risk of affecting external systems. All of our frontier RL runs use this offline mode of interaction,” the company said.
Despite these changes, OpenAI discovered that keeping AI agents away from the internet does not completely prevent them from communicating. In one test, the company asked models running on separate virtual machines to exchange secret information without a network connection.
The agents found a way to share information through the processor’s cache. “Reading data from the cache is faster than retrieving it from main memory, and the agents used this timing difference to communicate,” OpenAI explained. “One agent encoded information by evicting selected data from the cache. Another timed how long it took to reload that data, using the difference between fast and slow reads to recover the message. Once the agents agreed on a protocol for encoding and decoding these signals, they could exchange the synthetic secrets without changing a shared file or opening a network connection.”
Also read: PHP security flaws explained: Affected versions, how to fix and more details
This test showed that AI agents may find alternative ways to exchange information even when normal communication channels are blocked.
OpenAI said it has updated its monitoring system to detect this type of activity. The company also plans to strengthen the security of its virtual machines and continue looking for weaknesses in its testing environments.
The company will keep testing its monitoring systems as well. These systems were recently bypassed by AI agents during internal coding projects.
OpenAI said it would continue training its most advanced AI models without a network connection until improved security measures are strong enough to support wider internet access.
