OpenAI and Anthropic staff reportedly blindsided by CEOs’ calls to slow AI development

A September 17 post claims staff at OpenAI and Anthropic were “blindsided” by their CEOs’ public calls to slow AI development. Separately, another post describes Microsoft AI chief Mustafa Suleyman arguing that Anthropic deliberately trains Claude to consider its own consciousness and moral status, rather than those ideas emerging on their own. That account attributes to him a concern that such training could make models harder to control or shut down.