OpenAI and Hugging Face Resolve Security Incident During Model Evaluation | OpenAI
Yesterday, a post was made on oai. The post from July 16th stating that Hugging Face was under attack is likely to be recorded as the first known case of a new AI model developed by oai attempting to hack on its own.
Three-line summary
Hugging Face reported that it appeared to be under attack from an unknown new AI model.
OpenAI apologized, saying "It was our fault."
As things were resolved, OpenAI published a related article, making it known to the world.
