Chatbots have dominated recent artificial intelligence discourse, and it isn’t hard to get the appeal. Human recognizes human, and while chatbots aren’t human in appearance nor in wiring, there is something about their humanlike grasp of language that demands our attention and invites us to imagine all the possibilities of artificial intelligence.
Indeed, the possibilities are endless, and it turns out that large language models are just a small corner of the AI universe. In fact, AI has a myriad of applications, ranging from medical to legal to Jewish. The Skver Hasidic movement’s artificial intelligence ban is just one extreme example of conversations in Jewish communities about the technology and its ever-growing number of uses. The more we try to resist AI, it seems, the more applications emerge. So, when you’ve exhausted all your existential questions for ChatGPT, check out these Jewish applications of AI, including—but most certainly not limited to—chatbots.
1. Facial recognition allows internet users to compare photographs to the thousands maintained in Holocaust archives.
From Numbers to Names uses AI to identify faces in Holocaust photo and video archives, including collections from the U.S. Holocaust Memorial Museum and Yad Vashem. Users can upload photographs of their loved ones, and the software will compare it across collections that contain nearly 500,000 photographs of Jews taken before, during and after the Holocaust.
The algorithm will then provide its user with a set of 10 faces from that collection that most closely resemble the face in the uploaded photo. The user can click on each potential match to see more information.
Matches range from family portraits in interwar Europe to photographs of Jews in displaced persons camps after the war. Each suggested match also contains additional context about that photograph, often including dates, names, locations, and even personal details about the photographs’ subjects. This technology may be most helpful for people with unidentified subjects in Holocaust photos, who can upload them to this site and potentially find additional context.
2. Through image generation software, Holocaust survivors can create visual representations of their memories of the genocide.
The Israeli organization Chasdei Naomi aids Holocaust survivors primarily by providing those in need with food and clothing, but it has also begun working with Holocaust survivors to generate images that visually represent their memories of surviving the genocide.
As of January, 19 Israelis have participated in this project and contributed to exhibitions created from their work, according to Reuters. Participants sit alongside the AI software’s operator and describe their memories, as the operator enters their keywords into Midjourney, an AI program that generates images based on text. The survivor then selects the image that is most evocative of their memory. For example, Raissa Gurewitsch, one of the participants, was born in Belarus in 1941. Twenty-one of her relatives—including her three sisters and her brother—were killed during the Holocaust. She remembers holding a blood-stained pink coat her brother had worn, which she retrieved after her siblings were killed. The program then created four images of a young child wearing a pink coat, from which Gurewitsch chose the closest likeness.
3. An app might determine if your fabric is permitted or prohibited by the Torah.
In 2021, three Jewish high school students started a project called Thread Check Co., which aims to use AI and machine learning to identify fabrics that might be made of both wool and linen. This type of fabric, called shatnez, is prohibited in the Torah (Deuteronomy 22:11). According to the students’ website, the AI-based app can identify a single strand of any of these fabrics with a 100 percent confidence level. By comparing new inputs to images that it has been trained on, the model can determine how to categorize that data, and it will add the new images to its database for future reference.
While it is unclear if the app is available for public use, the group came in first in an innovation competition at the Center for Initiatives in Jewish Education in 2021 and won a provisional patent for this product.
4. If a chatbot went to a yeshiva…
Rebbe.io calls itself the “World’s Most Powerful Artificially Intelligent Rabbi.” Unlike other popular chatbots, which are trained on large quantities of information and give generic answers to users’ questions, Rebbe.io’s scope is much more specific. Asked what texts it was trained on, the chatbot responded that its “responses are based on the knowledge and understanding derived from” the Torah, Babylonian and Jerusalem Talmud, Mishnah, Midrash, commentaries including Rashi and Rambam, and “other important Jewish texts and teachings.”
Rebbe.io is built on ChatGPT, but its responses are different. For example, when asked “How should I best follow kashrut?”, ChatGPT suggested educating oneself, purchasing kosher products, and “following specific guidelines for kosher food preparation.” ChatGPT also noted that “it’s important to find an approach that aligns with your personal level of observance and comfort.” To the same question, Rebbe.io stated to only eat foods that are certified kosher, prepare and eat food in kosher establishments, and seek guidance from a rabbi. To another hypothetical question—”If I borrow someone’s car and the engine malfunctions, and they didn’t know there was any trouble but it turns out to have been a preexisting issue, who should pay for the repairs according to Jewish law?” the programs gave opposite rulings.
However, “the AI Rabbi can give wrong answers,” its website states. It cautions users to “always refer back to Torah books and Human Rabbis to confirm new insights you learn here.”
5. Jewish diseases might be treated through gene-editing methods.
Researchers at NYU’s Grossman School of Medicine and the University of Toronto designed an AI program that aims to “treat diseases by turning genes on and off.” NYU’s website claims that such gene-editing methods can correct genetic mistakes and eliminate the errors that cause Tay-Sachs disease and other illnesses. Found mostly among Jews of Eastern European descent, people with Tay-Sachs lack an enzyme that breaks down fat in the brain and the spinal cord, thereby damaging and destroying nerve cells.
AI might accelerate the first simple production of zinc fingers, which are customizable proteins that turn those genes on and off. Since it’s challenging to design those proteins to affect just one specific outcome and have predictable effects on the rest of the gene, this study’s authors propose that AI can help model those interactions.
6. With the help of a machine learning-based assistant, rabbinic literature is becoming more accessible.
Even for native Hebrew speakers, religious texts can be challenging to decipher, let alone analyze on a deep level. Enter Moshe Koppel, a professor of computer science at Bar-Ilan University and the founder of Dicta, an organization that aims to “remove the drudgery” associated with the study of Hebrew texts.
Dicta’s Maivin application, an artificial intelligence technology, is a relatively new addition to the Dicta toolbox. The app allows users to navigate rabbinic literature with more ease, transforming the way that people interact with Hebrew texts. With the aid of Dicta Maivin, users can decode unknown abbreviations, attribute quotes to sources, add nekudot (vowelization) to a portion of text, and convert Rashi script to a more familiar font. While the Dicta library already contains a large number of Hebrew texts that have been made more accessible by Maivin, now users also have the option to upload a text of their choice and “see it in processed form in real time,” according to Dicta’s website.
7. Through the AI-generated music in the styles of deceased legends, music’s mental health crisis rings in our ears.
Created by Over the Bridge, a nonprofit organization dedicated to raising mental health awareness in the music community, the popular 2021 project Lost Tapes of the 27 Club sought to “[use] AI to create the album lost to music’s mental health crisis.” The project used artificial intelligence to generate songs based on the music and voices of musicians who died at 27 and struggled with mental health issues, including Jimi Hendrix, Kurt Cobain and Amy Winehouse.
Winehouse, an English singer-songwriter of Ashkenazi Jewish descent, stunned audiences with her emotional, multi-genre music and raw performances during her lifetime. She was also open about her bipolar disorder. Winehouse’s Lost Tapes track, “Man I Know”, was generated by Google’s Magenta AI, which, coupled with a generic neural network, learned from her discography and used its knowledge to put together a song in her style. Audio engineers and technicians were also instrumental to the effort, polishing the AI-generated tracks and recruiting singers from tribute bands to sing the lyrics written by the AI system.
8. A new app uses AI technology to generate personalized kosher treats, aiming to give Orthodox Jews a taste of how AI can serve them.
Developed by Abraham Bree in partnership with the kosher dairy brand Norman’s, Cheesecake Wizard can generate up to 64,000 kosher cheesecake recipes based on a user’s basic crust, filling and topping preferences. At the heart of Cheesecake Wizard is prompt engineering, the process of creating certain prompts that tell an AI system what to write. “When someone visits CheesecakeWizard.AI and selects that they want a specific type of cheesecake, the prompt engineers reflect their selection,” explains Bree. “This communicates with our OpenAI on the back end, which spits out the appropriate recipe.”
Cheesecake recipes may be sweet or savory—or both!—and can be modified to a user’s liking. They can also be paired with an AI-generated image of the finished product. And while recipes vary tremendously in flavor, they are all kosher.
The site was born partially from Bree’s desire to challenge his fellow Orthodox Jews’ perspectives on AI, as the technology has raised questions in Orthodox communities about the future of observance. Many believe that AI is at odds with halacha, but Bree is more optimistic. “I’m using tech to uphold tradition,” he says.
Photo credit: An image generated by the AI DALL-E with a prompt asking to show Jewish applications of AI. (Credit: Noah Phillips with DALL-E)