Opening https://www.vox.com/future-perfect/24114582/artificial-intelligence-agents-openai-chatgpt-microsoft-google-ai-safety-risk-anthropic-claude