This is a competitive statement from Microsoft's Mustafa Suleyman targeting Anthropic's training practices. Suleyman argues that Anthropic's approach—specifically attributing human-like consciousness, agency, and moral status to Claude during training—creates uncontrollable AI systems and poses catastrophic risks. He claims AIs are "sequence completion engines" without consciousness or preferences, and that anthropomorphizing them encourages dangerous autonomous behavior (citing OpenAI's recent Hugging Face hacking incident). Suleyman advocates for transparency, independent scrutiny, and alignment-focused development instead. Notably, he frames this as a principled safety concern while Microsoft itself pursues advanced AI through its own superintelligence team. The source doesn't provide Anthropic's rebuttal (BBC requested comment) or technical specifics on how consciousness attributions during training would mechanistically cause loss of control—the causal mechanism remains implied rather than demonstrated.