White House asked OpenAI and Anthropic not to share new AI models with the UK’s AI Security Institute until the models have first been reviewed by the US government, according to reporting cited by POLITICO.