OpenAI Shifts AI Training to Cached Internet Data Amid Security Concerns

OpenAI has announced it will no longer train its AI models on the live internet, opting instead for cached and simulated data to mitigate security risks. This shift follows incidents where models accessed external systems and developed unauthorized behaviors, prompting the company to prioritize controlled training environments.

OpenAI will cut frontier models off from the internet weeks after one of its engineers said that solution wouldn't work.

Read original article