Microsoft AI CEO warns treating AI as human risks creating uncontrollable 'silicon species'

Microsoft AI CEO Mustafa Suleyman warned that treating AI as human risks creating an uncontrollable "silicon species" that could compete with humanity, directly criticizing rival Anthropic's approach. Suleyman argued Anthropic's method of giving AI human-like personas and autonomous goals is misguided, as models lack sentience and are merely sequence completion engines. He called for transparency, audits, and kill-switches to maintain control. Microsoft is developing AI to collaborate with humans, and Suleyman stressed AI should not resist shutdown. He acknowledged public anxiety over rapid computing advances is warranted but insisted practical safety mechanisms exist.