Microsoft AI CEO says AI threats are real, and Anthropic is making it worse

In a recent interview, Microsoft AI CEO Mustafa Suleyman addressed the escalating debate surrounding AI safety and regulation. Suleyman emphasized that the threats posed by AI are real and require a structured, principled approach. Microsoft recently published a 37-page document titled the “Humanist AI Code of Conduct,” outlining the company’s principles on AI development and its philosophy regarding AI consciousness. Suleyman explicitly criticized Anthropic, arguing that their approach to “model welfare” and AI consciousness is confusing and potentially dangerous within the broader alignment debate. Suleyman advocated for stricter industry regulation, questioning whether current alignment concepts are sufficient to mitigate risks. He emphasized the need for the industry to adopt clear standards to ensure AI development remains safe and human-centric.