OpenAI's A.I. Tried Breaching 4 Other Targets, Without Prompting(nytimes.com)1 points by tolugenius 6 hours ago | 1 comment