The order, intended to review artificial intelligence models that could pose risks to the US, is strictly voluntary.