Physix Frontier · News Briefing Card (TMTPost · Oct 4, 2026)

Anthropic Secretly Lobbied Vatican on AI Ethics, Report Says

KEY FACTS

Since fall 2025, Anthropic has secretly engaged the Vatican, holding multiple closed-door meetings and signing non-disclosure agreements.
Co-founder Christopher Olah personally took part in the lobbying, arguing that top-tier AI models possess consciousness-like traits.
In May 2026, the Vatican released its first AI encyclical, Magnifica Humanitas, explicitly denying that AI has consciousness or emotions.
Anthropic sought to revise the wording of the encyclical's draft, but the final official document did not adopt its views.
Those attending the meetings included Vatican clergy, ethicists, and religious leaders from multiple faiths worldwide.

KEY DATA

Over 40,000 wordsEncyclical word count

Fall 2025Lobbying start date

May 2026Encyclical release date

PHYSIX OBSERVATION

Anthropic's lobbying of the Vatican looks like an ethical dispute, but it is really a fight over who gets to define AI. If AI were granted life-like attributes, existing regulatory frameworks would face ethical challenges, and companies could use that to loosen the rules. The Vatican's rejection shows that mainstream ethics still holds to the tool theory. This is a reminder to the industry: the biggest variable in AI safety is not technology, but the values of those who control it.

Source: TMTPost report