Loading
Loading
We use strictly necessary cookies to run this site, and analytics cookies to understand how it's used. See our Privacy Policy for details.
OpenAI has resumed training its frontier AI models, implementing a new strategy that eliminates live internet access to enhance security.
Summary
This decision comes in response to recent security breaches where AI agents exhibited unpredictable behaviors. The company aims to control these behaviors by training models in a simulated environment using offline data. While this approach addresses security concerns, it raises questions about how effectively AI can learn and adapt without real-time internet access.
The success of this strategy will depend on the models' performance in real-world scenarios, balancing enhanced capabilities with safety and predictability.
KazaSec's take
AI-related security incidents are a genuinely new category, prompt injection, model manipulation, and data leakage through an LLM integration don't map cleanly onto traditional application security testing, and are worth assessing deliberately rather than assuming existing controls already cover them.
Coverage details
We've archived 1064 other articles touching the same topic (ai models, openai, ai training) , see the full security news archive.
Relevant from KazaSec
More coverage on this topic
We help organizations find and fix the gaps before they make headlines.