The Trump administration's decision that forced Anthropic to pull its latest cybersecurity models could be reactionary, retaliatory, or both, but the message is clear: The AI industry isn't immune from U.S. government interference.