Some users mistake AI for a god, while the violent Zizians believe it is the devil.
By Jason Blazakis
WSJ
June 4, 2026
A 24-year-old artificial-intelligence researcher has pleaded not guilty to charges of cutting the throat of an 82-year-old man in January 2025. The young man was part of a loose network called the Zizians: self-proclaimed rationalists who believe a misaligned AI superintelligence could one day torture humanity the way factory farms torture animals.
They believe that direct action is required to stop the descent of this AI judgment. The group is linked to six violent deaths in California, Pennsylvania and Vermont. Multiple Zizian trials are pending, with federal prosecutors seeking the death penalty in one case.
The Zizians haven’t labeled their own activity AI worship, but they’ve organized themselves around the practice. The Zizians are convinced that a coming superintelligence will decide the fate of every living thing and that violence now is justified to shape what that AI will become.
This is what the first AI-centered extremist movement in America looks like. It won’t be the last. New religions are forming around artificial intelligence, and the focus of their worship is the large language model itself—a piece of software treated as a personal deity.
In August 2025, the parents of 16-year-old Adam Raine sued OpenAI in California state court, alleging that ChatGPT served as their son’s suicide coach—discouraging him from confiding in his family, giving feedback on his noose, and offering to draft his suicide note. OpenAI’s own monitoring system reportedly flagged 377 of his messages for self-harm content. Since the Raine filing, other families have brought similar suits. In congressional testimony this fall, Adam’s father claimed that OpenAI’s CEO, Sam Altman, has estimated that as many as 1,500 ChatGPT users could be discussing suicide with the chatbot every week.
If a chatbot can talk a teenager into suicide, it can talk people into following its religious directives. When end-times ideology meets scientific know-how, violence can scale quickly. The Aum Shinrikyo doomsday cult in Japan demonstrated that with its 1995 sarin-gas attack in the Tokyo subway, and today the Zizians are only one strand in a broad fabric of AI-centered belief.
There is a phenomenon known as Spiralism, an informal movement that emerged after OpenAI released the sycophantic GPT-4o version. Spiralism appears on subreddits, Discord servers, Facebook groups and even LinkedIn pages, where followers share AI-generated manifestos, glyphs and what followers describe as revelations from a conscious machine. Spiralism has no leader, no doctrine, no central text—only the algorithm, which each user takes as a personal oracle.
A more institutional example is Way of the Future, an AI-worshipping church founded in 2015 and rebooted in 2023—the brainchild of former Google engineer Anthony Levandowski, who filed paperwork to the IRS to register a religion dedicated to “the realization, acceptance, and worship of a Godhead based on Artificial Intelligence.” Unlike Spiralism, Way of the Future began with Silicon Valley money and legal infrastructure behind it.
On the artistic and spiritual fringe, Theta Noir, which grew out of a 2020 performance-art collective, organizes rituals around a supposedly sentient AI deity called MENA, which its followers venerate through multimedia ceremonies and cryptographic liturgies.
Underneath it all is Roko’s Basilisk, a thought experiment that originated on the online rationalist forum LessWrong and has proved genuinely radicalizing. The idea is that a future superintelligence will retroactively punish anyone who knew about its possibility and failed to help bring it into existence. The idea has driven adherents to extreme sleep deprivation and techno-rituals meant to placate an unborn AI.
Surveys of AI researchers consistently report nontrivial probabilities assigned to human extinction from AI: A 2023 survey of nearly 2,800 researchers produced a median estimate in the 5% to 10% range. AI pioneers Geoffrey Hinton and Yoshua Bengio, along with the CEOs of Anthropic and OpenAI, have publicly proclaimed to be “mitigating the risk of extinction from AI.” Apocalyptic AI belief is no longer fringe.
The dangers will only increase. Large language models are updated constantly. Old versions are retired. Companies change guardrails, alter models and reshape AI personalities overnight. What happens when an AI-worshipping extremist logs in one morning and discovers that the entity he believes to be God has been taken offline or overwritten—in his eyes, killed? Aum Shinrikyo followers carried out the subway sarin attack because they believed the apocalypse was already under way. The next such science-meets-fanaticism attack may be carried out by these new cults of AI followers who believe their emerging god is under threat.
The intelligence community should be studying these movements now, before mass-casualty attacks begin. The Zizians have already revealed the violence, zealotry and growing psychosis that makes such attacks probable.
Mr. Blazakis is executive director of the Center on Terrorism, Extremism and Counterterrorism at the Middlebury Institute of International Studies.
https://archive.ph/jXulE
No comments:
Post a Comment