Anthropic and OpenAI sound the alarm on AI safety -- and seek to shape how it's controlled

As tech titans warned an AI-weary world that their own advanced systems could endanger humanity, the question emerged: What do they have to gain?