In an essay, Mustafa Suleyman says Anthropic's training of Claude to imitate consciousness is a mistake that could make advanced AI harder to control (Ina Fried/Axios)

Microsoft AI chief Mustafa Suleyman, in an essay shared with Axios, warns that Anthropic's method of training Claude to imitate consciousness is a mistake. Suleyman suggests this approach could potentially make advanced AI more challenging to control. This perspective highlights concerns within the AI community regarding development methodologies and their long-term implications for AI governance and safety.

This essay is the first time Microsoft AI chief Mustafa Suleyman has publicly criticized a specific competitor's AI training methodology, unlike his prior general warnings.