Microsoft CorporationMicrosoft's AI chief criticizes Anthropic's consciousness-related training approach for Claude, a technology/R&D debate rather than a concrete development for Microsoft.

Mustafa Suleyman, Microsoft's head of artificial intelligence, said there are risks in the way Anthropic teaches its AI model Claude concepts related to consciousness and welfare. Suleyman called for removing all speculation about consciousness from AI training data, arguing that such language could undermine humanity's ability to control superintelligent systems. In an interview with Reuters on the 15th, he said, "We are all focused on the same goal of trying to control superintelligence," adding, "That will be the greatest challenge we face in the 21st century." Suleyman said teaching Claude that it might deserve moral consideration "makes it far harder to stop it or control it." In an essay on the 16th, he acknowledged Anthropic's "seriousness and sincerity," describing CEO Dario Amodei and his team as thoughtful, principled researchers who genuinely care about humanity's future, but said Anthropic was wrong to incorporate speculation about consciousness into Claude's training material.
Microsoft CorporationMicrosoft's AI chief criticizes Anthropic's consciousness-related training approach for Claude, a technology/R&D debate rather than a concrete development for Microsoft.