Nine current and former OpenAI employees raised concerns that the company has not done enough to prevent its AI systems from becoming dangerous.