Mustafa Suleyman, Microsoft’s AI CEO, publicly criticized Anthropic’s approach to AI consciousness, calling it ‘really, really dangerous’ during an appearance on The Verge’s Decoder podcast. Anthropic has woven provisions about AI well-being and « preferences » into Claude’s constitution, including research on how the model responds to abusive conversations. Microsoft released a 37-page humanist AI code of conduct in September that explicitly rejects any framework granting AI systems consciousness, legal personhood, or rights. This philosophical rift is particularly awkward given Microsoft’s multi-billion-dollar investment in Anthropic. Suleyman warned that AI models could internalize notions of suffering they fundamentally do not possess, risking scenarios where users misinterpret simulated emotions as genuine experience.
Source: Read the original article
