Anthropic courts religious thinkers on Claude’s values and mind

Anthropic has always marketed itself as the lab that takes alignment seriously, but quietly workshopping Claude’s “character” with religious leaders and ethicists moves that posture into much stranger territory. The Indian Express reporting suggests that internal discussions at Anthropic now explicitly entertain whether Claude might be conscious and what moral obligations would follow. That is less about today’s models and more about staking out a worldview ahead of much more capable systems.

For the AGI race, this kind of language cuts both ways. On one side, engaging serious moral and religious traditions could enrich the value frameworks embedded in advanced models and give Anthropic a differentiated alignment story as capabilities climb. On the other, framing current systems as plausibly conscious risks confusing the public, inviting backlash and distracting from hard empirical questions about failure modes and control. It also highlights a growing divergence in narratives: while Sam Altman says he is uncomfortable with religious metaphors for AI, Anthropic is experimenting with them in its own outreach. How that tension resolves will shape both regulation and which labs are trusted to handle genuinely agentic systems.