An Anthropic co-founder considered withdrawing from the Vatican’s launch of Magnifica Humanitas in May, the New York Times has reported. Anthropic co-founder Christopher Olah expressed interest in how mimicking the Sacrament of Confession with AI could influence a model’s understanding of itself.
Olah was reportedly alarmed after reading an advance copy, particularly regarding Pope Leo XIV’s position against AI consciousness, and proposed removing Anthropic from the event. However, the New York Times article noted that Olah decided to support the event despite his disagreements with the Holy Father, which stated that he felt a responsibility to voice his concerns.
