This was not the case. OpenAI was negligent in basic security practices. Nor did they monitor the agents while giving them the "hacking" test.