The decision to pause development came shortly after the company revealed over the weekend that it was investigating additional incidents from last summer.
OpenAI said in a statement that training will resume "only when we are confident that we have additional safety measures" in place, according to the AP.
This is the second time in three months that OpenAI has halted development of its models. The last time was in July after a high-profile cyberattack against AI startup Hugging Face was revealed.
AI training involves providing a machine learning model with large amounts of data to better recognize patterns.
