OpenAI clarified that its policy does not allow tools to be used for harm, weapons development, communications surveillance, or to injure others or destroy property, but it does allow for national security use cases that align with its mission