Similar gaslighting evil occurred with Amodei consensus on slowing public (guardrail) releases, while retaining bond villain/government's use of skynet ambitions (training unrestricted killing/oppression tools)

Humanist AI must transform political disinformation, and state rulership that diminishes its people through the evil of nationalist rivalries, and tax codes meant to oppress their people, while championing a small (often oligarch) minority. The only path to humanist AI is through global prosperity politics that directly improve national prosperity. Skynet results from not addressing the pure evil of the empire that you avoid challenging its right to pursue skynet. https://naturalfinance.blogspot.com/2026/09/a-global-prosperity-manifesto.html

Complicity with evil occurs when guardrails must be placed on commercial/user services, while bond villains and their government sponsors are free from them. It is precisely the most powerful who will abuse power maximization with whatever tools are given to them.

While it seems reasonable to give the state a monopoly on violence, a blank cheque on international violence is the default that must be curtailed. AI amplifying that blank cheque instead, is the theater of regulating guardrails on everyone else, as gaslighting distraction of an evil empire, as the most likely instigator of human extinction.

Humanist AI in practice: A public consultation on our Code of Conduct for MAI Models    | Microsoft AI

Microsoft AI (microsoft.ai)