A voluntary arrangement for reviewing AI models was discussed at the White House earlier this week, reports Politico. OpenAI, Anthropic, Meta and Google were among companies attending the meeting to discuss the details of the arrangements, which were finally completed to meet a 1 August deadline. OpenAI, Anthropic and Meta had previously reviewed the draft arrangements and submitted edits to them in late July. It is widely thought that OpenAI and Anthropic have been asking for a compulsory review of products that would slow the import of increasingly capable models from China. There have been extensive objections to compulsory review from the US software industry, which sees OpenAI and Anthropic as promoting the case for compulsory review as a way to slow the competition.