According to the Washington Post, sources said that OpenAI rushed through safety tests before ensuring product safety.