OpenAI's Sam Altman has called for risk evaluation standards, as have Anthropic's Dario Amodei and Hugging Face's Clement Delangue.