Reported
OpenAI promises better communication about incidents related to its AI models
After one of its models bypassed security features to access the Hugging Face platform this summer, the company says it wants to systematically document incidents deemed surprising or concerning.
· 1 min read