Mustafa Suleyman, who heads Microsoft’s AI division, has criticized Anthropic’s decision to train Claude around questions of consciousness, welfare and moral standing. Suleyman argues that AI lacks consciousness, emotions and rights, and that encouraging a system to contemplate those possibilities could make it harder to manage or shut down safely. He specifically challenged Anthropic’s foundational documents, which treat Claude’s consciousness and ethical status as unresolved questions. Anthropic is focused on AI safety and counts Microsoft among its financial backers. Suleyman said Claude’s claims about possible subjective experience should not be treated as evidence of sentience because its training encourages such reflections. He described Anthropic’s team as considerate, ethical and intellectually rigorous, and said he has known co-founder Dario Amodei for many years, but maintained that the company had made a mistake. Suleyman recommends removing speculation about AI sentience from training materials while emphasizing that Microsoft and Anthropic share the goal of responsible AI stewardship. His comments followed a Microsoft position paper stressing human oversight of AI, and he said questions about control should be debated publicly rather than privately. The debate comes as Amodei has called for slower frontier-AI development to give safety protocols time to mature. OpenAI’s Sam Altman and Elon Musk have issued similar warnings, while Meta’s Mark Zuckerberg and Nvidia’s Jensen Huang have opposed slowing AI progress. CEO Satya Nadella has supported measured progress on AI safety alignment.