Fresh safety concerns follow OpenAI’s GPT-4 release

OpenAI's research paper reminded people to use GPT-4's outputs with "great care" because the AI model still "makes reasoning errors."