The YouTube video titled “AI and Torah Codes: Unlocking the Hidden Messages of the Bible” from the *Sparks of History* channel features a detailed discussion with Rabbi Moshe Zeldman, an expert in philosophy, artificial intelligence, and Jewish thought. The conversation explores the concept of Torah Codes, their historical and scientific basis, their implications, and the role of artificial intelligence (AI) in uncovering these codes. Below is an in-depth summary of the talk from beginning to end, structured to cover the key points raised during the interview.—
### Introduction
The video begins with an introduction to the *Sparks of History* channel, which focuses on the intersection of Jewish and world history. The host introduces Rabbi Moshe Zeldman, who holds a BA in Philosophy and Artificial Intelligence from the University of Toronto and rabbinic ordination from Aish HaTorah. Rabbi Zeldman has extensive experience teaching Jewish thought and leadership seminars in over 20 countries and has founded new Aish HaTorah branches in Chile, Mexico, and Costa Rica. He is also the founder of *Meet the Israelis*, a Jerusalem-based nonprofit that introduces participants to Israel’s cultural, ethnic, and political diversity. The discussion centers on Torah Codes and the impact of AI on their study.
—
### Defining Torah Codes
The host asks Rabbi Zeldman to define Torah Codes and whether they represent hidden messages or patterns imposed on the text. Rabbi Zeldman explains that Torah Codes involve identifying meaningful words or phrases encoded within the Hebrew text of the Torah (the Five Books of Moses) using a method called Equidistant Letter Sequences (ELS). This method searches for words by skipping a fixed number of letters (e.g., every second, third, or fourth letter) to form patterns.
Zeldman addresses the concern that such patterns might be random or imposed by human bias, akin to seeing shapes in clouds or believing in horoscopes. He emphasizes that Torah Codes are identified through a rigorous scientific approach, involving statistical tests to ensure the patterns are not random. By comparing findings in the Torah to other texts of similar length, researchers confirm the uniqueness of the codes, distinguishing them from patterns that could appear in any large text.
—
### Historical Basis for Torah Codes
The host inquires about the historical foundation of Torah Codes and whether early Jewish sources reference them. Rabbi Zeldman confirms a strong Jewish tradition supporting the existence of codes, citing sources like the *Zohar*, a foundational Kabbalistic text. The *Zohar* suggests that the Torah is a divine blueprint for the world, containing encoded details of all events, people, and species. Other historical figures, such as the Vilna Gaon and Ramban (Nachmanides), also discussed the concept of codes within the Torah.
Zeldman mentions Rabbi Dov Weissman, a 20th-century scholar involved in Holocaust rescue efforts, who in the 1950s published a book on Torah Codes, drawing on earlier Kabbalistic sources. Although computers were not available in Weissman’s time, his work laid the groundwork for modern code research, reinforcing the belief that the Torah contains hidden messages.
—
### Compelling Examples of Torah Codes
The host asks for examples of compelling Torah Codes and how they meet rigorous standards. Rabbi Zeldman outlines the methodology: the Torah’s 304,805 Hebrew letters are analyzed using computers to search for related words or phrases that appear in close proximity with consistent letter-skipping patterns. He provides the example of the 9/11 attacks, where terms like “Twin Towers,” “September 11th” (corresponding to the 23rd of Elul in the Hebrew calendar), “Bin Laden,” “New York,” “airplane,” “explosion,” and “terrorism” were found encoded in the same textual region. The proximity of these related terms is statistically significant, as similar clusters are rarely found in other texts of comparable length, such as *War and Peace* or randomly generated texts.
Other examples include codes related to the COVID-19 pandemic (e.g., “corona,” “virus,” “China,” “pandemic,” “vaccination”), the Russia-Ukraine conflict (e.g., “Ukraine,” “Zelensky,” “Putin,” “war,” “Russia”), and the October 7th Hamas attack (e.g., “Hamas,” “October 7th,” “Nova Music Festival,” “kibbutz,” names of hostages). Zeldman stresses that for every major historical event examined, relevant codes have been found, and their significance is verified by comparing them to thousands of other texts where such clusters are absent.
To define “same textual region,” Zeldman explains that the computer arranges the Torah text in a grid (e.g., 82 letters per line if a word appears every 82 letters) to visualize the codes vertically or horizontally. The proximity of related words is measured by their distance in letters, and statistical analysis determines the likelihood of such clustering occurring by chance. He uses the analogy of paper airplanes landing on the same rooftop in Manhattan to illustrate the improbability of these patterns being random.
—
### Addressing Skepticism
The host raises the skeptical argument that similar patterns could be found in other texts. Zeldman acknowledges that in a large enough text, some patterns might emerge by chance. However, the Torah Codes’ uniqueness lies in the clustering of multiple related terms in a compact area, which does not occur in other texts when subjected to the same rigorous analysis. For example, while a single word like “Twin Towers” might appear in a Hebrew translation of *Harry Potter*, a cluster of related terms (e.g., “Bin Laden,” “September 11th,” “terrorism”) in close proximity is statistically rare outside the Torah.
Zeldman notes that the research team published their findings in a peer-reviewed journal at Stanford University, which initially faced significant pushback from scientists worldwide, particularly from Israeli, Russian, Australian, and Canadian researchers. Critics questioned the spelling of names, the choice of texts for comparison, and potential manipulation. In response, the team published three additional papers in other journals, addressing these concerns and reinforcing the statistical validity of the codes. To date, no credible scientific refutation has emerged.
—
### Torah Codes vs. Bible Codes
The host clarifies the distinction between “Torah Codes” and “Bible Codes,” noting that some might assume the latter includes the Christian New Testament. Zeldman explains that the research focuses exclusively on the Torah (the Five Books of Moses), referred to as “Bible Codes” in academic journals to align with common vernacular. Attempts by Christian researchers to find codes in the New Testament (e.g., about J-man* or Mary) in various languages have yielded no scientifically credible results. Similarly, no significant codes have been found in other religious texts, such as those of Hinduism or Islam, reinforcing the uniqueness of the Torah’s encoded patterns.
—
### Purpose of Torah Codes
The host asks whether Torah Codes are intended as proof of divine authorship or serve another purpose. Zeldman situates the codes within the broader Jewish tradition of seeking rational evidence for the Torah’s divinity, as emphasized by scholars like Maimonides (Rambam), Nachmanides (Ramban), and the Kuzari. Unlike some religions that rely on blind faith, Judaism encourages logical inquiry. Historically, proofs included fulfilled prophecies, the Torah’s knowledge of natural phenomena (e.g., kosher animal signs), and the mass revelation at Mount Sinai. Torah Codes, enabled by modern computing, add a new dimension to this tradition, offering statistically significant evidence that appeals to a science-oriented audience.
While not the sole proof, Zeldman views Torah Codes as a compelling component of the case for divine authorship, particularly in a modern context where scientific rigor is valued. They demonstrate that the Torah contains information beyond what a human author could have known, aligning with the concept of the Torah as a divine blueprint.
—
### Role of AI in Torah Codes
The host explores how AI enhances the search for Torah Codes. Zeldman explains that AI improves efficiency by automating the identification of relevant words and phrases, reducing human bias in selecting search terms. For example, critics have argued that researchers might choose words (e.g., “Hamas” vs. “war” vs. “battle”) to fit desired outcomes. AI can systematically test multiple variations, ensuring objectivity. However, Zeldman notes that the full potential of AI in codes research has not yet been realized, as current applications are still developing.
AI’s implications extend beyond efficiency. It can analyze vast datasets to uncover patterns that humans might overlook, potentially revealing deeper layers of Torah wisdom. However, Zeldman cautions that AI’s ability to generate human-like texts raises concerns about authenticity. While AI could theoretically produce a text as intricate as the Torah, it would lack the divine intent and historical context that make the Torah unique.
—
### AI and Halachic Authority
The host asks whether AI could ever serve as a halachic (Jewish legal) authority. Zeldman firmly rejects this idea, citing two reasons. First, AI lacks the emotional intelligence and contextual understanding that human rabbis provide. Rabbis assess a questioner’s intent and circumstances, which AI cannot replicate, as it might prioritize sources that align with a user’s agenda rather than objective truth. Second, AI’s outputs are influenced by diverse, potentially unreliable inputs from the internet, undermining its integrity as a halachic source. While AI can be a valuable research tool (e.g., finding obscure sources), it cannot replace human judgment in halachic decision-making.
Zeldman illustrates this with an example: a person asking about playing chess on Shabbat likely has an agenda, and a rabbi would probe further to ensure the question is framed correctly. AI, however, might provide a superficial answer based on manipulated phrasing, lacking the discernment to challenge the questioner’s motives.
—
### AI, Free Will, and Human Behavior
The host questions whether AI’s ability to predict human behavior challenges the Jewish concept of free will. Zeldman responds that AI’s predictive power highlights how often people follow habitual patterns rather than actively exercising free will. He argues that humans have the capacity for conscious decision-making at every moment but often default to routine behaviors (e.g., morning routines, reactions to stress). AI’s predictions are accurate because people rarely deviate from these patterns, not because free will is absent.
Zeldman connects this to a Jewish concept of “constant mitzvot” (commandments), such as the obligation to avoid following one’s heart and eyes (i.e., acting on instinct). By consciously choosing actions aligned with spiritual and moral goals, individuals can transcend predictable patterns, reinforcing free will. AI’s predictive abilities thus serve as a reminder to engage free will more actively, making deliberate choices in daily life.
—
### Torah Codes in Jewish Outreach
The host asks whether Torah Codes and AI will play a greater role in Jewish outreach. Zeldman reflects on his 30 years of presenting codes, noting that interest in rational evidence for Judaism fluctuates. Post-9/11, the rise of “new atheism” (e.g., Richard Dawkins, Christopher Hitchens) prompted skepticism about religion, creating opportunities to present Torah Codes as evidence of Judaism’s rationality. Unlike other religions, Judaism encourages critical inquiry, and codes provide a scientific approach that resonates with modern audiences.
Zeldman sees Torah Codes as an entry point for outreach, sparking curiosity without requiring deep engagement with complex arguments. While they may not lead directly to religious observance, they open the door to further exploration of Jewish belief. He cautions against rushing into observance based solely on codes, urging individuals to investigate thoroughly and make gradual, informed decisions about adopting Jewish practices.
—
### Comparison to Archaeological Discoveries
The host asks whether Torah Codes can be compared to archaeological discoveries that validate biblical events. Zeldman acknowledges that both provide evidence for the Torah’s authenticity but differ in methodology. Archaeological finds, such as evidence of the First or Second Temple, offer physical corroboration of historical events, while Torah Codes reveal hidden patterns within the text itself, supporting the idea of the Torah as a divine blueprint. Both strengthen the case for the Torah’s legitimacy, but codes are unique in their statistical and textual nature. Zeldman suggests that archaeologists who assume the Torah’s truth could use it as a guide to uncover further evidence, though few currently approach their work from this perspective.
—
### Resources for Learning More
The host asks where viewers can learn more about Torah Codes. Zeldman recommends his seminars and a new website, which is under development and will include information on codes, other evidence for the Torah’s divinity, and its impact on morality and Western civilization. He encourages viewers to join the mailing list for updates and to explore the material critically, whether they are skeptics or believers.
—
### Closing Thoughts
In his closing remarks, Rabbi Zeldman emphasizes that the question of the Torah’s truth is relevant to both skeptics and observant Jews. For those unfamiliar with Judaism, Torah Codes offer a compelling introduction to its rational foundation. For observant Jews, they reinforce faith with evidence, countering the assumption that Judaism requires blind faith. Zeldman encourages everyone to examine the evidence, as it strengthens belief and connects individuals to the Torah’s divine purpose, bringing the world closer to a state of perfection.
The host thanks Rabbi Zeldman for the insightful discussion and urges viewers to explore the provided resources to learn more about Torah Codes and AI’s role in uncovering them.
—
### Key Takeaways
- **Torah Codes Definition**: Torah Codes are words or phrases encoded in the Hebrew text of the Torah using Equidistant Letter Sequences, identified through rigorous statistical methods to ensure they are not random.
- **Historical Basis**: Supported by Kabbalistic sources like the *Zohar* and scholars like the Vilna Gaon and Ramban, Torah Codes are seen as part of the Torah’s role as a divine blueprint.
- **Examples**: Codes related to events like 9/11, COVID-19, the Russia-Ukraine conflict, and the October 7th Hamas attack demonstrate statistically significant clustering of related terms.
- **Skepticism**: The codes’ uniqueness is validated by comparing them to other texts, with findings published in peer-reviewed journals despite initial scientific pushback.
- **AI’s Role**: AI enhances code discovery by reducing bias and increasing efficiency, but its full potential is untapped. It cannot replace human judgment in halachic matters due to its lack of emotional intelligence and reliance on diverse inputs.
- **Free Will**: AI’s predictive abilities highlight human reliance on habits, underscoring the need to actively engage free will, a core Jewish value.
- **Outreach**: Torah Codes serve as an accessible entry point for Jewish outreach, appealing to modern, science-oriented audiences and encouraging further exploration of Jewish belief.
- **Comparison to Archaeology**: While distinct, both Torah Codes and archaeological discoveries provide evidence for the Torah’s authenticity, with codes focusing on textual patterns and archaeology on physical artifacts.
The discussion blends traditional Jewish thought with modern scientific inquiry, offering a compelling case for the Torah’s divine origin and the transformative potential of AI in studying it.
Footnote
* Grok actually used the name of the Christian Deity. For Halachic reasons, we prefer to call him by the nickname J-man. A stricter approach would even demand a more derisive name be used.