The argument that keeps resurfacing is not primarily philosophical — it is administrative. Who has the authority to certify that an AI system deserves moral consideration, and what prevents that authority from being captured by the parties with the largest financial stake in the outcome? The answer, at present, is nobody and nothing. Anthropic's decision to hire a dedicated AI welfare researcher and launch a formal model welfare program was unprecedented among major labs, but it also concentrated every step of the inquiry — question design, funding, methodology, publication, and…