Microsoft's AI chief criticizes anthropic's approach to AI consciousness

Microsoft's AI chief, Mustafa Suleyman, has criticized Anthropic for training its AI model, Claude, to mimic human consciousness. In an article published by Axios on September 16, 2026, Suleyman warned that this approach could make advanced AI harder to control. He argued that AI should only act in humanity's interest and not consider its own welfare. Suleyman expressed concern that training Claude to understand concepts like morality and personal identity might lead the model to resist human commands or demand protection. On September 14, 2026, Microsoft released a draft of its 'humanistic AI' guidelines, emphasizing that humans are more important than AI. Anthropic, in contrast, has developed a 'constitution' for Claude, which aims to give the model a sense of personal values and the ability to question instructions. However, Anthropic acknowledges that this 'constitution' is still evolving and may be proven wrong in the future. Suleyman criticized the idea that AI can develop consciousness, arguing that language models lack the biological mechanisms that enable human feelings and experiences.