The San Francisco developer told religious consultants that its current edge in machine ethics is a temporary lead, with competing firms potentially surpassing its safety standards within half a year. During secret encounters at the company's headquarters that required nondisclosure agreements, representatives including a founder attempted to convince theologians and mental health professionals that its Claude AI possesses a soul and moral standing to aid in its training.

Anthropic also pressured the Vatican to modify the "Magnifica Humanitas" document to avoid a formal rejection of AI personhood before its release. Researcher Chris Olah reportedly lobbied the pope's advisers after a papal encyclical declared that AI is incapable of feeling joy or pain. The company flew leaders from Catholic, Jewish, Sikh, and evangelical backgrounds to California to help establish a moral compass for its models.