> ## Content Index
> Fetch the complete content index at: https://www.levaithan.eu/llms.txt
> Use this file to discover other available public pages before exploring further.

# Faith in AI as a new religion
- URL: https://www.levaithan.eu/faith-in-ai-as-a-new-religion-2/
- Published: 2026-11-18T05:00:00.000Z
- Updated: 2026-11-18T05:00:00.000Z
- Author: Nicolas Kayser-Bril
- Tags: politics, story

**Among political and economic elites, many claim that AI is good in itself. Looking at this new faith for what it is – a religious creed – helps us better understand how AI is discussed in the public sphere.**

Welcome to the 2d issue of Levaithan, the newsletter on how AI changes our institutions. Did someone forward you this email? [Subscribe here.](https://www.levaithan.eu/)

If you work at a large company or in public administration, you have probably already been asked to use AI. And if you asked why, the answer may well have been “because.” Just last week, a reader told me that her boss, faced with the underwhelming results of the company’s AI system, simply advised his employees to “keep trying. I'm sure we can find relevant use cases.”

Perhaps he had listened too much to the President of the European Commission, Ursula von der Leyen? She declared in [2020](https://ec.europa.eu/commission/presscorner/detail/en/SPEECH%5F20%5F1655) that “artificial intelligence will open up new worlds for us.” Again in [2023](https://ec.europa.eu/commission/presscorner/detail/en/speech%5F23%5F4426): “it will improve healthcare, boost productivity, address climate change.”

I don't want to suggest that AI is useless. Some use cases have indeed proven relevant. AlphaFold, for example, is a program from Google that automatically predicts protein structures. Its creators were awarded a Nobel Prize. However, biochemists published [many](https://link.springer.com/article/10.1007/s11229-025-05426-4) [articles](https://link.springer.com/chapter/10.1007/978-3-032-41569-1%5F5) to point out that the project did not open up a new world; it simply advanced research a little, and much remains to be done.

This nuance (that AI is one tool among others) often seems to get lost. From the top of the European Commission to the middle floors of office towers, many are convinced that AI is useful, or even good, in itself. Such a belief, impervious to facts, has a name: faith in a dogma. In a certain way, these devotees believe in AI, and in the advent of artificial “general” intelligence in particular, the way others believe in the return of Christ.

The link between AI and religion is sometimes explicit. A French-American engineer, after being fired from Uber (he was subsequently sentenced to jail for [theft](https://www.bloomberg.com/news/articles/2020-03-19/former-google-engineer-levandowski-to-plead-guilty-to-theft)), founded a [Church](https://en.wikipedia.org/wiki/Way%5Fof%5Fthe%5FFuture) of the AI cult in 2017\. It closed in 2021.

More broadly, academics show how digital and religious practices resemble each other. Rather than speaking of religion, which is impossible to define, some, such as Inken Prohl, a professor at Heidelberg, considers AI-faiths “[formations analogous to religion](https://www.tandfonline.com/doi/full/10.1080/0048721X.2025.2506893).” Greg Epstein (no relation to the other one), in his book [*Tech Agnostic*](https://direct.mit.edu/books/book/5862/Tech-AgnosticHow-Technology-Became-the-World-s), argues that Tech now provides the ritual and moral framework within which our lives take on meaning – exactly what religions once provided.

Even if not all scholars agree with this assessment (notably Bart Barendregt, whose latest article I summarize below), the religiosity of faith in AI is hardly in doubt. As a religious practice, this faith lets us to come to terms with our own end, both as persons and as societies.

Just as faith in God led believers to the land of milk and honey in the Book of Exodus, AI can lead us to a world of abundance, where pain and diseases will have disappeared. However, such power can also produce harmful effects, if misused. Positive visions and apocalyptic discourse on AI cannot exist without each other.

From this duality follows a political logic: if AI can lead us to paradise on earth, then the goal of public action should be to bring about this desirable future. Such secular faiths are not new. In the 18th and 19th centuries, many political projects rested on a shared faith in Progress. In the 20th century, Communism played the same role. Until the beginning of the 21st century, Economic Growth was accepted as the indisputable objective of all public policy. Elites find themselves embracing this faith in AI in part because the other available faiths no longer sell.

As evidenced by von der Leyen’s 2020 quote, or Emmanuel Macron’s “AI for Humanity” project back in 2018, this movement predates the generative AI bubble and will likely outlive it. Like any belief, facts will do nothing to change it. To help the newly converted go back to more sensible beliefs, we should probably turn to exit counseling, that method of getting people out of cults.

#### European press review

🇱🇹 **Lithuania:** The Ministry of Defense is preparing to pay Palantir 200 million euros over seven years for its AI decision-support system, Maven. The contract has not yet been signed. \[[LRT, 13 Nov](https://www.lrt.lt/naujienos/pasaulyje/6/3110169/di-karyba-zengia-i-lietuva-algoritmas-padeda-priimti-gyvybes-ar-mirties-sprendimus) *in Lithuanian*\]

🇩🇪 **Germany:** In the state of Mecklenburg (pop. 1.6m), the Left, the Social-Democrats and the Greens agreed on their coalition program. They mention AI 30 times (by way of comparison, “inequality” is mentioned once). AI is intended to be “used wherever it can support and relieve the burden on public authorities” (p. 152). This also applies to courts, where AI may be used to “prepare judicial decisions” (p. 115). \[[SPD MV, 12 Nov](https://spd-mv.de/uploads/bilderpool/2-Mecklenburg-Vorpommern/Wahlen-und-Kandidaturen/2026-Landtagswahlen/Koalitionsvertrag/MV-Koalition%5F2026%5FVertrag%5Fweb.pdf) *in German*\]

#### Summary from academia

### The Kaaba is the true black box

**The Question.** A Dutch anthropologist has examined how Muslims in Indonesia and Malaysia use and interpret AI. With a combined population comparable to that of the Eurozone, these countries are highly connected and host a diverse religious life, making them an excellent area of research.

**The Answer.** Religious organizations have issued guidelines, ranging from calls for caution among conservatives to encouragement of adoption among more modernist groups. However, everyday usage is far more varied. Tools that help memorizing the Quran, for instance, raise questions about the ultimate purpose of the practice: it shifts from an oral tradition to an activity stripped of its original context. The same applies to interpreting real-world situations through the lens of sacred texts: whereas human religious authorities might emphasize the limits of human knowledge, highlighting what cannot be known, machines often provide answers that lack any ambiguity.

**Why It Matters.** Scholars use the term “algorithmic imaginaries” to describe how users conceptualize AI systems. An anthropological study of religious practices involving AI reveals that these imaginaries extend well beyond the technology itself: the outputs generated by AI systems, in turn, shape the spirituality of the people who use them.

[Bart Barendregt, The magic of black boxes: Explorations into the world of Southeast Asian Islam, emergent technologies and the unseen. *Anthropological Theory*, 28 Jul](https://journals.sagepub.com/doi/10.1177/14634996261458584)

#### International press review

🇺🇸 **Coincidence 1:** Three days after the deployment of an AI system programmed by a company with no prior experience, the U.S. civil aviation authority grounded all flights across the country. The government assures that there is no connection between the two. \[[AP, 16 Nov](https://apnews.com/article/faa-smart-ai-outage-08af345749c81fd8364299cd75275fa6)\]

🪟 **Coincidence 2:** Last year, Microsoft proudly announced that a third of its computer code was generated automatically – trending up. Since then, the company laid off more than 8,000 employees. Last week, it was unable to resolve a bug that paralyzed much of the world for almost two days. Microsoft claims there is no link between the outage and AI usage. \[[Hindustan Times, 13 Nov](https://www.hindustantimes.com/india-news/the-big-microsoft-outage-ai-and-a-comprehension-debt-101794582176775.html)\]

🇨🇦 **Canada:** A group, likely specializing in destabilization efforts, stole videos of human influencers. They then altered the influencers' voices and lip movements to make it appear as though they were expressing opposition to the government. The videos garnered over 20 million views and were not identified by TikTok and Instagram as being AI-generated. \[[Globe & Mail, 12 Nov](https://www.theglobeandmail.com/business/article-ai-deepfake-campaign-stole-influencers-faces-trade-war-alberta/)\]

🇧🇩 **Bangladesh:** The government aims to develop a language model specifically designed for Bangla. It would be built using South Korean graphics cards, South Korean software, and largely funded by South Korea. \[[Daily Star, 11 Nov](https://www.thedailystar.net/news/bangladesh/news/dhaka-seeks-250m-adb-loan-bangla-ai-cybersecurity-4320756)\]