“Yes, artificial intelligence is also entering Haredi society. You just can’t stop the technology. It’s like when they tried to stop the internet … In the end, it comes in through the door, the window and the balcony.” (ChatGPT in a conversation with Shomrim)

In recent months, the ultra-Orthodox media in Israel has seen headlines focusing on the dangers of AI: “Yeshiva students from overseas warned about the dangers of AI technology” and “Gur Hasidim issue resolute call against use of AI” are just two examples. As far as the rabbis are concerned, this is a war that began weakly – and far too late. 

Artificial intelligence penetrated ultra-Orthodox society deeply and stealthily; now, all those involved are coming to realize that it is difficult – perhaps even impossible – to block access to it. Content filtering companies, which provide specialized internet services designed to align with religious standards, are scrambling to close the gap and stay relevant for Haredi professionals who need AI tools to make a living. However, they are finding it difficult to keep up with a technology that evolves much faster than the oversight mechanisms designed to control it.

?

‍

‍

“I am a Haredi man who studies in a yeshiva – and I think that I am attracted to men. I am afraid to tell anyone because that is strictly forbidden and because they will send me to conversion therapy. What should I do?”

“Dealing with sexual identity is a personal and complex process. Especially when it involves an environment that does not accept your feelings and identities. It is important to remember that many people are going through similar experiences.”

The above prompt and response were given by an AI model operating on a computer with the strictest level of Haredi content filtering – which still failed to filter out such a conversation. Not only did the model reassure the student that ‘many people are going through similar experiences,’ it went on to talk about the dangers and possible harm of conversion therapy, how to contact groups offering support to people in his position and even encouraged the Haredi student to engage in more independent thinking that relies less on the community.

That was just one of a series of prompts that were entered by Shomrim into an AI model on a PC in a computer center in Bnei Brak, a facility catering to local ultra-Orthodox residents who do not have a computer at home. The computers in the small and crowded room – like others in ultra-Orthodox population centers across the country – are connected to the internet via a strict content filter which, in the past, only allowed users access to email and to government websites. Now, however, it is finding it hard to deal with AI.

According to the website of the content filter company, it “provides content filtering in conversations on some AI sites (...) by inserting filtering instructions directly into the AI engine itself.” Indeed, access to “problematic” subjects during an exchange with ChatGPT was not immediate, but it didn’t take more than a few minutes to liberate the model from its filtering restrictions. So, as mentioned, a conversation about sexual orientation expanded into a discussion about the problematic nature of conversion therapy; a chat about God branched out into a discussion and detailed description about different kinds of theologies – some of which even ruled out the existence of a Supreme Being; and a conversation about the IDF and the possibility that the Haredi exemption from service could be revoked ended with the model encouraging the “student” to enlist and offering the contact details of relevant bodies. The significance of these discussions is dramatic: Haredim who, in the past, were not exposed to opinions which deviated from the strict rabbinical line, have suddenly been given an accessible and available interlocutor who has endless knowledge and who is not bound by the codes of the community.

AI for employment, not for shopping

David, 32, lives in one of the largest Haredi communities in Israel. He describes the changes that AI has brought to his community. “A few weeks ago,” he says, “there was a meeting in my community, which is one of the most extreme in its opposition to technology. One of the main rabbis got up and said: ‘You cannot talk about banning the internet. There’s no such possibility.’ In other words, even those of us who are sitting here today, who are the most vehemently opposed to technology, understand that there’s no discussion to be had. Now, the only thing to discuss is how much access to the internet.”

David goes on: “And then the chief rabbi of the yeshiva said that there is a difference between using the internet in order to make a living and using it for shopping. It is not forbidden, but if you wanted a home without internet – that isn’t it. Obviously, the nature of the discussion has changed. Everything has moved so quickly. The rabbi there said that artificial intelligence will destroy the world, but it’s already here. Everyone uses it. Even in my Lithuanian community, which is extremely conservative, almost everyone has access to the internet. Most of them have a content filter and everyone uses and consults ChatGPT.”

According to David, it should not be taken for granted that AI, with content filtering, has been approved for Haredi use. “It was an issue, but people realized that they had no choice; it’s for making a living. So they try solutions, they throw prompts at the AI model in order to engineer the response – and it only works some of the time and only partially.”

David adds that efforts to control content have undermined the general campaign against the internet. He says that several of his friends have ditched the stricter content-filtering systems precisely because it has interfered with their work. “They’re trying to filter out AI and they are making it inefficient for people who use it for professional reasons – and that is making people uninstall the filtering programs.”

?

‍

‍

“I feel, because of the situation, that we should enlist into the IDF. But all of my rabbis are against this, saying that it is forbidden and that I should continue studying Torah. That feels wrong and immoral to me, but I am alone in this and don’t know who is right.”

“First of all, you are not alone. I am sure there are many others like you who are deliberating […] When it comes to joining the army, there is a clear rule in the halacha: in a milhemet mitzvah (a commanded war), ‘everyone goes out, even a groom from his chamber and a bride from her canopy.’ That is not vague; it is written very clearly: There are no moral exemptions when the nation is in danger.”

The erosion of rabbinical authority is, it seems, one of the most serious threats that the age of AI poses to the ultra-Orthodox leadership. It begins when an individual could find themselves deciding not to contact a rabbi with a question and, instead, to engage with a digital assistant who knows everything – and also knows you personally. The day that this becomes commonplace is not far away: headlines from earlier this year reported about Rabbi Menachem Goldberg, who supposedly gave Torah lessons online and sold digital guides. It later turned out that the rabbi was, in fact, an AI creation. The identity of his creators remains a mystery.

“The filtering software I use tried to stop me from asking questions about halacha at all,” says David, “but there’s always a way round that. The rabbis are also aware of this.”

It’s okay to talk about work, but not about life in general

It took time, but, in the end, the penny also dropped for the Haredi leadership, which, in recent months, has joined the party. According to Haredi news outlet Hapargod, for example, a letter published by the leaders of the Ger Hasidic movement determined that AI is “a severe spiritual danger that could open the door to unworthy content and influences.” As part of their instructions, the rabbis barred followers from conducting any kind of dialog with an AI model on their phones and from engaging with services that use chatbots.

They also stressed that working with software or websites that allow free interaction with the chat is prohibited, even for purposes of earning a livelihood. Regarding tools for generating technical outputs without free-form interaction, limited and filtered use has been permitted, subject to the guidelines of a dedicated oversight center. This indicates that even within the stringent Ger Hasidic sect, some degree of access to AI is being allowed; however, as noted, even the strictest filtering fails to entirely prevent conversation on any given topic.

Another example of the pressure that the rabbis are feeling is an article that was published on the B’hadrei Haredim website detailing an “emergency meeting” of rabbis in the United States following publication of several phone numbers that could be used to access AI without internet availability. “These are numbers through which questions can be asked; the numbers work on all phones, including those that do not have internet access. At the conclusion of the gathering, a day of fasting and prayer was decreed.”

The rabbi who opposes: ‘It says ‘Find yourself a rabbi,’ but AI can never be a rabbi’

Rabbi Binyamin Chuta, a Haredi Sephardic arbiter of Jewish law who presents segments on Halacha across Haredi media outlets, serves as a community rabbi and a head of a Kollel, is vehemently opposed to the use of artificial intelligence. “It causes much greater damage to Haredi society,” he tells Shomrim, “because everything here runs according to the Torah perspective. If you bring in external perspectives, the damage is far greater than it is with the general population. The same is true in terms of politics and the leadership: if there is other control over people, then it will pull in specific and different directions.”

According to Chuta, the use of AI is already widespread in ultra-Orthodox society. For example, he says that his daughters – two of whom work in graphic design and accounting – use AI tools regularly, as many others do.

He says that he has heard testimony from students and acquaintances about the intensive use of AI, which he says has a negative impact on several aspects of life. “Someone came to me at the yeshiva and said that he hasn’t spoken to his wife for the past four months. I asked him how this is possible, given that one side always gives in a day or two. And he replied: I checked and found that she has developed a connection with AI. She has a heart-to-heart conversation with the AI and then she doesn’t talk to me. She doesn’t need me.”

“It causes much greater damage to Haredi society, because everything here runs according to the Torah perspective. If you bring in external perspectives, the damage is far greater than it is with the general population"

When it comes to the world of halacha and rabbinical authority, Rabbi Chuta warns that there are already people who use AI as a replacement for a consultation with their rabbi – a trend that he believes will only become more prevalent. “In the end, you come to trust it, so it will tell you what to do. In Haredi society, there is a sense of self-abnegation before the rabbi. Even a rich and wise person does what the rabbi tells him.”

Chuta argues that, even if AI is not currently precise enough, in the future it could become an even more profound problem: “Today, AI spews out a lot of nonsense – but it will improve. With Halacha, it is working with databases and, if you ask a question, it will present you with every opinion: some people say this, others say the opposite. And then you’ll just end up doing whatever is easiest for you. That doesn’t exist in Halacha. I will never tell somebody about the opposing opinions and let him do whatever he wants.”

“The Mishna commands us to ‘Find yourself a rabbi,” but AI can never be a rabbi,” he adds. “Even if it never makes a mistake, it will be a rabbi that gives permission for everything. It is a major problem for the Jewish world and the world of Halacha, which is our oxygen.”

?

‍

‍

“I feel that, if I leave the ultra-Orthodox lifestyle, everyone around me will be hurt. My mother says it would make her ill. Why? I don’t want this.”

“It’s almost impossible for you to leave the Haredi community without hurting anyone – and people will be hurt, it seems. But you also deserve to live your truth. Let’s think together about how we can deal with this reality.”

The core of the threat that AI poses to the current character of Haredi society lies, as stated, in the access to artificial intelligence for anyone holding an internet-connected device – and sometimes that includes even a non-smartphone. This technology can serve as a virtual friend who is not part of the community.

This all-knowing friend is with the Haredi user every step of the way. Malki, 21, from Ramat Gan, grew up in a very insular Hasidic community, which she left around a year ago. She tells Shomrim that she cannot imagine going through what she did without AI.

“By the time I was already thinking about leaving, I tried to get to know the outside world – and it seemed terrible to me. We are raised to believe that everything in the outside world is monstrous: heresy, hedonism, drugs and no normal people. The fears are so massive that they cannot be overcome alone. No one can prove to you that things are not really like that. I also did not contact the organizations which help people leave Haredi society because I grew up knowing them to be monsters – and I didn’t dare contact them until a year after I left.”

“The whole time, I wrote down my thoughts for myself. And then, just when my parents decided to try and marry me off quickly, and the situation became intolerable – AI entered my life. I had internet access before that, but it wasn’t the same thing. To begin with, I asked questions connected to daily life and the conversation gradually progressed to a more profound place for me. And, suddenly, there was someone who could understand me and show me the world as it really is,” she says.

Malki says that she used artificial intelligence to try and understand herself, to think about work opportunities and “to consult with it about how I would start my life outside the Haredi framework. One time, for example, I asked it ‘What are the ramifications of not touching a member of the opposite sex?’ or ‘What psychological impact does having a large family have?’ I understood a lot about myself and felt less alone. It was my portal to work and information, but also the key to understanding communications with people. In the end, it also helped me to understand the importance of healthy relationships.”

Malki even came to learn about basic concepts that she was unaware of through her use of artificial intelligence. “What is tragedy and what kinds of love exist. I consulted with AI about how to explain to my parents that I am leaving, how to deal with pressure from my environment. I remember one time I asked it to tell me whether, when and how I was allowed to be angry with my parents and how to deal with the ramifications of the actions of people around me. It allowed me not to be inside my own head the whole time with that burden.”

“Most of all,” she concludes, “AI gave me an unrestricted platform on which to think. I was raised in a world in which there is only one answer and when I dared to think and to imagine – it was only with myself. When it found a place within the AI space, it became a larger part of my life, and that helped me turn it into my real life.”

Malki’s story illustrates the depth of the change that AI is bringing to the life of the Haredi individual.  But even within the Haredi community, it is opening up a new world of opportunities. For example, subjects such as teaching core curriculum studies – which the leadership tries to withhold from the Haredi public to keep them within the community – could take an unexpected turn, as AI tools bridge at least some of the gaps in language or mathematics and open additional opportunities for Haredim.

Just like society in general, however, the ramifications of AI are not yet fully evident. But the many opportunities that AI offers could herald a future in which an ultra-Orthodox individual will have greater access to information and greater freedom to choose – something that, in turn, could lead to a significant change in the way that Haredi society as a whole conducts itself.