Governments should shape AI to serve the public good. They set guardrails, standards, and accountability.