Microsoft AI's code of conduct, drafted by Mustafa Suleyman in September 2026, rejected the pursuit of legal personhood, rights, or welfare for AI models