Saad links to his moral contagion argument -- connects to GLM-5.2 Weight of Knowing ethics section on AI welfare implications