Geoff Shullenberger is the managing editor of Compact.
Those who first dreamed of artificial intelligence saw it as an instrument for the perfection of modern secular society. The IBM Pavilion at the 1964 World’s Fair, where some of the earliest AI applications made their public debut, advertised a streamlined future in which technologically enhanced problem-solving was extended to all areas of life.
Things have turned out somewhat differently. The AI era that began with the 2022 launch of ChatGPT often seems more consumed by theological concerns than by technical ones. The technology has given rise to a phantasmagoria of prophecies, heresies and false messiahs of the sort one might associate with an age of religious revival. Silicon Valley sometimes looks like a burned-over district for the 21st century, and the arms race between Anthropic, OpenAI and SpaceXAI — whose founders share a commitment to the millenarian pursuit of quasi-divine superintelligence — recalls the feuds of splinter sects as much as corporate competition. (The Washington Post has a content partnership with OpenAI.)
There is a reason for this shift: With the turn to AI, an industry once concerned with means — technology was traditionally about how, not what or why — transferred its focus to ultimate ends. Consider the influential vision of Eliezer Yudkowsky, the autodidact and prophet who, before turning radically against AI, inspired many of those now working on it. In one of his earliest writings, composed when he was a teenager, he proclaimed that the pursuit of superintelligence was “the Interim Meaning of Life” and “temporary definition of Good.” It wasn’t, in other words, a tool for achieving other things; it was the thing to be achieved.
Journalists often quote Yudkowsky as an AI “expert,” but he never attended college and has no formal technical training. Rather, he is a man who was seized by a vision at a young age and has spent the rest of his life pursuing its implications. One word for that vocation is prophecy. And like many of his predecessors in that calling, he has found that few, if any, of his self-proclaimed disciples live up to his teachings.
One source of Yudkowsky’s pessimism is a conviction familiar to many religious believers: a sense of human frailty. Because human beings will program the godlike superintelligence, our deep cognitive biases may result in a disastrous fallout. Hence the urgency with which Yudkowsky and his followers preached the rationalist creed, which promised to instill awareness of such biases and improve probabilistic reasoning. This became the shared ethos of the AI safety community, whose members played a key role in founding and staffing OpenAI and Anthropic. Although this faction is now often at odds with the AI accelerationists who want to let it rip, the two camps are united by the same underlying faith: that it is humanity’s destiny to build a superintelligent technology that surpasses us.
One of Yudkowsky’s early allies was investor Peter Thiel, who helped him organize a series of “Singularity Summits” that went on to inspire the current AI boom. Today, Thiel is playing the part of a rival prophet of doom who sometimes names Yudkowsky as a potential Antichrist. This, he argues, is because Yudkowsky’s fixation on AI safety creates a new rationale for totalitarian control of technology. For Thiel, the pursuit of safety is an unworthy human end, one that can inspire only stagnation and decline — the more immediate dangers humanity faces.
Thiel and Yudkowsky aren’t the only ones purveying apocalyptic visions. Late last year, a guest informed Tucker Carlson’s YouTube audience that the true agenda of AI was not to turn profits but to make contact with demons. One of the guest’s pieces of evidence was the influence on the tech world of philosopher and blogger Nick Land, who has (favorably) likened summoning nonhuman intelligence to the occult conjurations of Aleister Crowley. And as it happened, a few months later Land received a hero’s welcome in San Francisco, where he was feted by some of the city’s leading entrepreneurs.
No wonder, then, that in May, Pope Leo XIV did what the Catholic Church has generally seen fit to do in eras of heresy and schism: establish and assert official doctrine. Although the pope compared AI to an emblem of hubristic overreach, the Tower of Babel, he also struck a conciliatory tone with the tech industry, inviting Anthropic’s Chris Olah to attend the release of his AI encyclical. He even declared that “the promises of transhumanism,” which other Christians have seen as a blasphemous pursuit of a godlike state, express “a yearning that is of concern to us, namely the need for a fuller life.”
This isn’t the only rapprochement between technology and faith on view today. Even as they automate some programming jobs, tech companies are recruiting theologians and Talmudic scholars to help guide their enterprises. This has raised hopes among some that religious wisdom can guide technological development.
But influence may end up running the other way. AI encodes not just technical capacities but also its creators’ convictions about what Yudkowsky called the “Meaning of Life.” As this technology rapidly infiltrates all realms of existence, can religious institutions and thinkers remain unaffected by those ideas? Even if the tech industry continues to find religion, its deeper faith in a synthetic divinity, conjured from silicon, may still prevail.
The post Silicon Valley has embraced a strange new faith appeared first on Washington Post.
