Judaism and Artificial Intelligence: From the Golem to GPT

The Oldest AI Story

Long before Silicon Valley, there was Prague. According to legend, in the late 16th century, Rabbi Judah Loew — the Maharal of Prague — fashioned a figure from clay, inscribed the word emet (truth) on its forehead, and brought it to life. The Golem was created to protect the Jewish community from blood libels and pogroms. It was powerful, obedient, and tireless. It was also, eventually, dangerous — growing beyond its creator’s control until Rabbi Loew had to deactivate it by erasing the first letter of emet, leaving met (death).

The Golem story is not just folklore. It is a thought experiment about the creation of artificial entities — their uses, their risks, and the moral responsibilities of their creators. And it predates ChatGPT by about 450 years.

Judaism has been thinking about artificial intelligence — in the broadest sense — for a very long time. The tradition’s resources for engaging with AI are richer, stranger, and more relevant than most people realize.

The Talmudic Precedent

The Golem is not only a legend. The Talmud (Sanhedrin 65b) records that the sage Rava created a man — a kind of artificial human — and sent it to Rabbi Zeira. Rabbi Zeira spoke to the creature, and when it did not respond, he declared: “You are from the companions” (meaning: you were made by the sages, not by God). He then told it to “return to your dust.”

This brief passage has generated centuries of commentary. What exactly did Rava create? Was it truly alive? Did it have consciousness? Could it speak? Could it sin? Could it be held responsible for its actions?

These questions — framed in terms of clay and mystical incantations — map remarkably well onto contemporary debates about AI. Is an AI system that passes the Turing test “alive” in any meaningful sense? Does it have moral status? Can it be held accountable? The language is different; the questions are the same.

Can AI Create Torah?

One of the most provocative questions in contemporary Jewish thought: can an artificial intelligence write a valid Torah commentary?

On one level, AI can already do this. Language models can generate coherent, even insightful interpretations of biblical texts. They can draw connections between passages, identify literary patterns, and produce divrei Torah (words of Torah) that sound like they could come from a human scholar.

But Jewish tradition has always held that Torah study is not merely an intellectual exercise — it is a spiritual encounter. The act of studying Torah is itself a mitzvah, an act of connecting with the divine. The Talmud speaks of Torah scholars as partners with God in the ongoing revelation of meaning. Can a machine be a partner with God? Can an algorithm have kavvanah (intention)?

Most authorities would say no — but the question forces a deeper inquiry: what exactly makes human Torah study sacred? Is it the content produced or the consciousness of the one producing it? If an AI generates an insight that moves a human being and brings them closer to God, does the AI’s lack of consciousness diminish the insight’s value?

These are not hypothetical questions. Rabbis are already grappling with them as AI-powered Torah study tools become available.

Shabbat and Automation

Shabbat — the weekly day of rest — prohibits melakha (creative labor). This includes actions like kindling fire, writing, building, and many others. But what about actions performed by a machine that was set up before Shabbat?

Jewish law has long dealt with automation. The Shabbat clock (timer) is widely used to turn lights on and off automatically. Slow cookers keep food warm without human intervention. Elevators in Israel are programmed to stop at every floor on Shabbat so that no one needs to press a button.

AI introduces new complexities. A Shabbat timer is simple — it follows a preset schedule. An AI system, by contrast, makes decisions. It responds to conditions, adapts, and takes actions that its programmer did not explicitly specify. Is an AI-driven thermostat that adjusts based on occupancy patterns the same as a simple timer? What about an AI security system that decides whether to alert authorities?

Some contemporary halakhic thinkers argue that as long as the AI system is operating within parameters established before Shabbat, it is analogous to a timer — a form of grama (indirect causation), which is generally permitted. Others are more cautious, arguing that the decision-making capacity of AI makes it qualitatively different from a timer and requires fresh halakhic analysis.

Pikuach Nefesh: AI and Saving Lives

The principle of pikuach nefesh — that saving human life overrides almost all other commandments — provides a powerful Jewish framework for evaluating AI in medicine.

AI systems can already detect certain cancers more accurately than human radiologists. They can predict cardiac events, identify drug interactions, and assist in surgeries. From a Jewish ethical perspective, if an AI system can save lives, there may be an obligation to use it — just as there is an obligation to use any available medical technology.

But pikuach nefesh cuts both ways. If an AI system makes a medical error — misdiagnoses a condition, recommends the wrong treatment — who is responsible? The doctor who relied on it? The company that built it? The hospital that deployed it? Jewish law has a developed framework for medical liability (rofe), but it assumes a human decision-maker. AI complicates this framework significantly.

Autonomous Weapons

The ethical challenges become starker when AI moves from the hospital to the battlefield. The development of autonomous weapons — systems that can identify and engage targets without human intervention — raises profound questions that Jewish ethics is well-equipped to address.

The Torah establishes rules of war that include proportionality, distinction between combatants and civilians, and the obligation to offer peace before attacking (Deuteronomy 20). These rules assume a human moral agent making decisions in real time. Can an algorithm satisfy these requirements? Can a machine distinguish between a combatant and a civilian with the nuance that Jewish law demands?

Rabbi Yuval Cherlow, an Israeli ethicist, has argued that if autonomous weapons can reduce civilian casualties compared to human soldiers (who are subject to fear, rage, and confusion), there might be a moral case for their use. Others counter that delegating the power of life and death to a machine is a fundamental violation of human dignity — tzelem Elohim, the image of God in which every person is created.

The Image of God Question

Underlying many of these debates is a foundational theological question: what does it mean to be created in the image of God (b’tzelem Elohim)?

If the divine image resides in consciousness, creativity, and moral agency, then AI — no matter how sophisticated — lacks tzelem Elohim and can never have the moral status of a human being. It is a tool, however complex.

If the divine image is expressed through the act of creation itself — through making something new, something that did not exist before — then humans creating AI are exercising their most God-like capacity. The creation of AI is not playing God; it is being the image of God.

Both interpretations have support in Jewish tradition. And the tension between them is productive. Judaism has never been a tradition that resolves its deepest questions with simple answers. It holds contradictions in creative tension, and the question of AI is no exception.

A Living Tradition

What makes the Jewish conversation about AI distinctive is that it is not starting from scratch. Judaism has 3,000 years of thinking about creation, consciousness, responsibility, and the boundaries between the human and the divine. It has the Golem. It has the Talmudic sages who created artificial beings. It has a legal system — halakha — that has adapted to every technological revolution from the printing press to the automobile to the internet.

AI will be no different. The questions are new; the framework for asking them is ancient. And in a tradition that has always believed that the most important questions are the ones without easy answers, the age of artificial intelligence is not a crisis. It is an invitation — to think harder, argue more, and bring the best of Jewish wisdom to bear on the most consequential technology humanity has ever created.

The Golem awaits. The question, as always, is what we write on its forehead.