The moves come when OpenAI’s chatbots, which can engage in human-like conversations and create videos and images based on text prompts, have become increasingly powerful and have stirred safety concerns.
OpenAI assigns new project to AI safety leader Madry in revamp
