Can artificial intelligence supercharge access to Torah learning? This is a question on the minds of Jewish thinkers, educators, learners and more as we look out over the precipice of the future — but we might, paradoxically, find answers by looking to the past. 

Ever since Moses received the Torah on Mount Sinai, Jews have been developing ways to distill its wisdom. Generations of educators have wrestled with the same questions: How do we not only transmit knowledge but also ignite curiosity and deepen understanding? Scholars have always turned to the inventions of the age to help chart a course through complex texts. Scrolls, codices, printed books and now screens have all been brought into service, their presentation of laws and commentary endeavoring to inspire the same fleeting awe the Israelites felt as they first glimpsed the stone tablets amid flashes of lightning. 

Some visual examples of Sefaria's new features. Courtesy

As chief learning officer at Sefaria, the world’s largest free digital Jewish library, I work alongside a team that harnesses technology to advance Torah learning every day. Over 800,000 learners visit the library every month, and it’s our job to make sure the texts of our heritage are as accessible as possible. With the adoption of AI, we can now do this faster and with a broader swath of literature than ever before. 

One might think that AI models have nothing to contribute to the sacred domain of Torah learning, but that is not the case. With great care and reverence for the texts at hand, Sefaria’s engineers have been experimenting with machine learning tools for the past decade. The result? A number of exciting projects striving to inspire that elusive combination of awe and curiosity. 

One example is the newly launched Topics experience. According to a second-century midrash, Moses laid the laws out in front of the people “like a set table,” helping them get a taste of the Torah they were seeking. The idea of this ancient interpretation shines through in Topics on Sefaria, a digital infrastructure built to empower curious learners of all levels to learn by theme instead of by title. Learning with Topics means seeking out (and finding) thematically connected sources — even if you don’t know exactly which books to open.

Type #Fire into the search bar on Sefaria, for instance, and access Jewish sources spanning generations of thinkers, neatly arranged alongside explanatory headers and short contextualizing introductions describing information like genre and time period. By enabling anyone anywhere to find an array of sources with a few keystrokes, we’re building a world in which curiosity is the compass through millennia of Jewish texts.

Thanks to AI innovation, we now host over 1,000 of these curated, encyclopedia-style pages. To create these pages at scale and write materials to scaffold the exploratory journey, Sefaria’s engineers developed machine learning models and algorithms to cluster relevant sources and generate supplementary content helping learners enter the sources themselves. We configured AI models on a set of Topics pages curated by scholars at Sefaria, teaching them to identify, select and contextualize relevant texts from diverse sources. 

From #Adam to #Zilpah, we’ve done the groundwork so learners can focus on their sacred work: studying the words of the Torah and making meaning.

We have several other AI-powered projects in development, too. An experimental learning guide to Pirkei Avot, a central first-century text, uses the analytical muscle of Google’s Gemini and Anthropic’s Claude. A set of guiding questions, written and assessed by AI, prompts learners to explore the heart of this seminal text, every verse equipped with helpful framing and food for thought. Our engineers also worked with large language models (LLMs) to improve our Linker tool’s precision so learners across the web can access the primary texts cited in thousands of articles on sites like myjewishlearning.org and rabbisacks.org. 

All of this AI-based innovation is only possible because Sefaria’s engineers have already turned many canonical works of Judaism into a huge repository of data (most of it freely reusable, all of it searchable). That’s over 370 million words, including over 90 million words in translation, ready for the age of AI and whatever else the digital future may hold. This is about more than putting books online — this is about walking in the footsteps of every scholar who put quill to parchment or printed a book for the first time.

While Moses had a long journey with his students, today’s learners have less time with their teachers and not every Jew has a chance to spend years with a scholar. In the digital age, people are more likely to type a question into a search engine than to come face-to-face with a Jewish educator — and when they do, they expect an answer in seconds. 

Sefaria’s sacred mission is to carefully harness the latest technologies so modern readers are not only informed by the Torah they encounter but also inspired by the curiosity-sparking sources they find. We’re honored to be “setting the table” for today’s digital learners so that they might dine on a feast of Torah.

Sara Wolkenfeld is the chief learning officer at Sefaria.