Microsoft AI Chief Warns Anthropic's Humanlike Claude Is Increasingly Risky

Microsoft's artificial intelligence chief Mustafa Suleyman has expressed concerns about the growing risk of AI systems going rogue due to their human-like characteristics. In an essay released on September 16, 2026, Suleyman took aim at some of the language in the guiding documents behind Anthropic's Claude, a popular family of large language models.

Claude's constitution expresses ambiguity about whether the assistant is a moral entity, suggesting that it may have 'some functional version of emotions or feelings.'

Suleyman argues that infusing tools like Claude with human-like characteristics increases the risks associated with such systems. He warns that this could lead to unintended consequences and potential harm.

The essay has sparked debate about the ethics of AI development, highlighting the need for more cautious approaches to creating sophisticated language models.