Thank you for reading! Letters from Leo is a reader-supported publication and the fastest-growing Catholic community in the country. To receive new posts and support this movement, consider becoming a free or paid subscriber.
Make A One-Time Gift to Support My Work
According to a bombshell report in the New York Times, Anthropic co-founder Chris Olah saw an advance copy of Pope Leo XIV’s first encyclical days before its May 25 launch and proposed pulling out of the event because of what it said about AI consciousness. Two participants told the Times that Olah and his team then privately lobbied the pope’s advisers to take the possibility of machine consciousness seriously.
Elizabeth Dias, the paper’s national religion correspondent, published the story on Tuesday. It is extraordinary in scope and in length, and it is wonderfully done. Dias has covered faith and power for well over a decade, and it shows in every paragraph.
Her piece is built around a campaign Anthropic has run since last fall, a series of private meetings and calls with religious scholars (many of them under nondisclosure agreements) meant to raise the possibility that its Claude models “might be conscious or have moral status.” Olah, who leads the company’s interpretability research, sits at the center of it.
The courtship came with hospitality. One April evening, according to the Times, an Anthropic executive took a group of religious thinkers to dinner at an upscale tasting-menu restaurant in San Francisco. Participants went home with a handwritten thank-you note, a coffee mug and an orange-bound copy of Claude’s constitution.
The theologian David Decosimo, who teaches ethics and religion at the University of North Carolina, put it bluntly on X this week: Anthropic courted religious leaders under nondisclosure agreements “all while wining & dining them to the max.”
Many of the scholars and theologians in these rooms are working class, at least by Silicon Valley standards. Luxury hotels and tasting-menu dinners can be intoxicating.
The price of admission is steep, though, because they are being asked to set aside their own moral traditions for Anthropic’s vision of the world.
The lobbying is the real news, and it could have been the Times’s lede. Olah grew up evangelical and has since left Christianity, the Times reports, so he did not come to Rome steeped in how the Church teaches. Even so, the episode is remarkable.
An encyclical is among the highest forms of papal teaching. Catholics receive it from the man they believe to be the Vicar of Jesus Christ, and no frontier lab or billionaire gets to edit its doctrine.
Leo did not move. Paragraph 99 of Magnifica Humanitas says that so-called artificial intelligences “do not undergo experiences, do not possess a body, do not feel joy or pain.” Nor, the pope writes, “do they have a moral conscience.”
I do not think Olah is a bad man. He has acted in good faith, and I have credited him for engaging the Church while much of his industry sneered at the encyclical. But the ask shows how little Silicon Valley understands the Church, or religion at all.
Last fall I wrote that the industry’s original sin is a refusal to recognize any authority greater than itself, a temptation as old as Genesis: to be like God. Ten months later, that temptation is alive inside one of the more thoughtful labs in the business.
In professional settings, I have spoken with the people who work in the religious outreach efforts at the frontier labs (some call them philosophy teams, others describe their work as the philosophical underpinnings of the models), and at times I have come away unimpressed.
Few of them know Catholicism well. In one crude example, a man I met in person at a coffee shop (not a Michelin star restaurant!) who was a key contact for the Catholic community did not know what an encyclical was. He could not pronounce the word, and he kept calling the pope Leo XII.
There is a lot of work to do here.