Unnoticed for a week!! The most worrying thing is that OpenAI ends up publicly acknowledging that they do not know how to contain a dangerous model but at the same time minimizing the incident and using it as marketing. The passivity with which the news was taken in general is
This shameful incident being used as marketing.
Basically OpenAI is telling the world that they don’t know how to test their AI models in a secure way. What next? Sorry we blew out a nuclear reactor while testing but we will take more precautions next time.
we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this.
openai.com/index/hugging-…
