Posts

Showing posts from November, 2026

Programmed to please: the moral and epistemic harms of AI sycophancy