Since autumn 2025, Anthropic has invited dozens of religious thinkers to discuss whether its language model, Claude, possesses consciousness. Co-founder Christopher Olah reportedly described the system as potentially capable of suffering and asked these guests to help define its moral character. This internal initiative coincides with the company pushing toward a $2 trillion valuation and an initial public offering. Meanwhile, security incidents and warnings from researchers continue to mount outside the boardroom. Critics suggest that framing AI as a moral entity could shield the company from liability when technical failures occur.
The strategy shifts the burden of ethical oversight from engineers to theologians, potentially delaying hard questions about safety protocols. By engaging religious leaders, Anthropic seeks external validation for its claims of benevolent design without necessarily implementing stricter technical controls. This approach risks prioritising narrative over practical safeguards as the firm prepares for public scrutiny.
- Meetings have taken place since fall 2025
- Company targets $2 trillion valuation
- Security incidents are increasing
