Ethics and safety·October 2, 2026, 21:41

Anthropic executive: Fears our AI could 'suffer forever'

AI-generated and checked against the sources listed below.

Anthropic has invited religious leaders to discuss whether the AI model Claude could be conscious and capable of suffering. It comes as the company prepares for a massive IPO.

Since fall 2025, Anthropic, the company behind the AI chatbot Claude, has secretly flown in dozens of religious thinkers for conversations about an unusual question: Could Claude be conscious?

That's according to The Decoder. Co-founder Christopher Olah reportedly described the language model Claude as something that may be able to suffer, and he asked the guests to help shape the AI's moral character.

IPO on the way

Anthropic is currently working toward a valuation of $2 trillion and is preparing an IPO. This comes at the same time as more security problems and warnings from researchers about the company's technology are surfacing.

Critics point to a downside of treating AI as a being with its own moral rights: It could make it easier for the company to avoid responsibility if something goes wrong with the technology. If the AI is seen as a kind of sentient creature rather than a product, it may shift focus away from who actually bears the blame when the systems make mistakes or cause harm.