Q&A: Charles Camosy, a Catholic theologian, discusses his experiences participating in meetings with Anthropic, which were convened to help shape the morality of its powerful AIs.
According to recent reporting by The New York Times, artificial intelligence (AI) juggernaut Anthropic spent much of this year picking the brains of various religious leaders — including Jews, Sikhs, and Catholics — on questions of “goodness” and morality.
The goal of the meetings was ostensibly to help Anthropic learn how to apply “centuries of human moral wisdom” to its AI models and products. These AI models cannot be forced to act morally, Anthropic says, but instead must be taught how to do so. The conversations played out amid fears — including those expressed by Anthropic representatives — that AI may one day surpass humans in intelligence and escape humanity’s control.
In addition, Anthropic’s leaders seemed eager to broach the topic of whether AI could ever become conscious, although strong disagreements persisted. The New York Times reported that Anthropic co-founder Chris Olah nearly pulled out of appearing next to Pope Leo XIV at the Vatican launch of Magnifica Humanitas over the encyclical’s strong rejection of AI consciousness, although he ultimately attended.
The Register spoke
