By Willow (willowashmaple.sbs)

Religious and moral education for artificial intelligence?

September 14, 2026

This past weekend, one of the top news headlines was about a former Anthropic researcher sounding the alarm (1) on the rapid development of AI, claiming that AI could drive humanity extinct by the end of this decade.

Albert Mohler today discussed this issue in his podcast, The Briefing (2). Mohler notes that this type of ontological question has arisen in the past, for example, when nuclear weapons were invented, and even in the infancy of modern computing. A Christian worldview Mohler espouses makes it clear that AI is a creation of humans; it does not possess consciousness, and it "is not a species. It is not a being. It is impersonating a being. It may be very convincing in impersonating a being, but it is not a being."

Nonetheless, AI is a mirror of human intelligence, except it is faster and is prone to magnify the worst human instinct it has learned, while also liable to errors and hallucinations unique to AI.

Back in April, I wrote (3):

Technologists have raised concerns over the rapidly improving capacity of generative AI and its potential for significant harm. For AI models, which are soulless algorithms that merely mirror human thinking, empathy is impossible. But can religious moral precepts be taught to AI as a kind of guardrail? For me, religion functioned as a powerful guardrail that kept me from destructive behavior. Ancient human civilizations needed religion to redirect primal animal instincts toward building a coherent and functioning society. Maybe artificial intelligence, being a mirror (albeit limited) of human brains, can’t it be trained on moral and ethical precepts derived from traditional world religions?

With the latest debates on AI and its dangerous potential, I am curious if experiments should be conducted in which AI is subjected to various types of moral and ethical education.

For example, one AI model is trained in Confucianism, while another receives traditional Christian moral precepts. Another AI model could be trained using modern humanistic ethical education and social-emotional learning. There could also be an AI model taught by Hassidic Jews, and a model that is trained by imams and Islamic ethicists.

If AI is a simulation of human learning and logic, such an experiment could also help study the development of moral reasoning and compare how different religions influence and shape the human concept of morality and ethics.

C.S. Lewis famously said, "education without values, as useful as it is, seems rather to make man a more clever devil." Perhaps training AI with values may also help stop humans from unleashing a modern-day Golem.

---

(1) Ex-Anthropic researcher: AI could kill us all within years

(2) AI and the end of humanity (September 14, 2026)

(3) Musing on artificial intelligence (April 27, 2026)

----

Use of this site's content is subject to the terms of Cooperative Nonviolent Public License (CNPL) ver. 7+.
A summary of license terms.

return home
headlines
subscribe (rss)
tip jar