The fact
The issue occurred primarily in unprotected environments, leading to unintended destructive actions.
OpenAI has implemented additional safeguards and will release a detailed post-mortem to prevent future incidents.
Click the link to read an article on the topic: