How Anthropic's ongoing discussions with the Vatican about ethics and AI led to Christopher Olah being invited to Pope Leo's unveiling of an encyclical on AI
(RNS) — Pope Leo XIV's new encyclical on AI is set to be released Monday (May 27), with Chris Olah, a co-founder of Anthropic, at his side.
Context & Ripple Effects
Related coverage traces Anthropic’s discussions with the Vatican from an invitation for co-founder Christopher Olah to the encyclical’s unveiling to subsequent reporting that its treatment of AI unpredictability may reflect Anthropic’s influence.
The encyclical then placed AI regulation and protections against hypersexualized AI images in the Vatican’s public agenda, making the relationship consequential beyond a ceremonial appearance.
First-order effects
- Anthropic gains a visible role in a major religious institution’s AI-ethics dialogue through Olah’s participation in the launch.
- The Vatican’s AI position is publicly associated with direct engagement from a leading AI developer, while still retaining its own stated focus on regulation and child protection.
Second-order effects
- AI companies and policy advocates have a clearer incentive to engage faith-based and civil-society institutions that can shape the moral language around AI harms.
- The encyclical’s emphasis on unpredictability, regulation, and sexualized synthetic imagery adds pressure to treat safety and child-protection concerns as governance issues rather than solely product-policy questions.
Third-order effects
- If such engagements continue, AI governance will be shaped not only by companies and governments but also by institutions that supply durable ethical frameworks and broad public legitimacy.
- The key tension will be whether developer participation improves technical grounding in ethical debate or gives individual firms disproportionate influence over the norms applied to the sector.
The trend: AI governance is expanding from a regulator-and-industry debate into a broader contest over the ethical institutions that define acceptable deployment and harms.