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

This newspaper opens in a new window