Introducing AskHillel - a Jewish ethical chatbot ~ Elder Of Ziyon

Wednesday, May 14, 2025

Introducing AskHillel - a Jewish ethical chatbot

Hillel the Elder was famous for his ethical sayings:

What is hateful to you, do not do to your fellow.
.
If I am not for myself, who will be for me? And if I am only for myself, what am I? And if not now, when?

Do not judge your fellow until you have stood in his place.

Be of the disciples of Aaron, loving peace and pursuing peace, loving mankind and bringing them closer to the Torah.

In a place where there are no men, strive to be a man.

He also was known for his extreme patience with everyone, famously failing to lose his cool when two people made a bet that they could upset him with inane questions.

Hillel the Elder is the model for my Jewish ethical chatbot.

As my readers have seen, over the past couple of months I have focused a great deal on Jewish ethics. I believe that the Jewish ethical framework is the most mature, practical, flexible and moral ethical system there is.

In the course of writing about it, I found that no one ever defined Jewish ethics in a structured way as separate from halacha (Jewish legal thought) and as a universal moral framework that applies to everyone, Jewish or not, believer or not. So I came up with a framework that is based on halachic ethical principles but without the Jewish-specific components that can more than hold its own against any other ethical system from Buddhism to utilitarianism to deontological ethics.

I realized that once I defined the rules, meta-rules and axioms that underpin Jewish ethics, I could create an AI chatbot that could adhere to those rules. The results have been most gratifying, and the chatbot is outperforming my expectations.

AskHillel is designed as a tool for exploring and applying a structured Jewish ethical framework to real-world dilemmas, debates, and decisions. AskHillel is not a halachic authority and does not provide legal rulings, but it draws deeply from Jewish moral values, philosophical traditions, and ethical reasoning to guide thoughtful reflection and moral clarity.

Unlike most chatbots, AskHillel uses a Dynamic Context Interpreter, which means it asks clarifying questions if your question has hidden assumptions or missing background. This mimics traditional Jewish debate and encourages deeper thought. it helps you define your question and intent, often uncovering your own biases before the question is even addressed. This one feature makes AskHillel a better tool in many ways than most general purpose chatbots.

It operates from a structured ethical system grounded in Jewish values such as Pikuach Nefesh (value of life), Kavod HaBriyot (human dignity), Emet (truth), and Anavah (humility.) These values are balanced using various triage rules which helps resolve conflicts between core principles in political or societal dilemmas.

AskHillel's guiding axioms include that objective truth exists, morality matters, and people can grow. Many modern ethical systems reject one or all of these axioms, and Jewish ethics can counter these malign yet popular ethics frameworks.

Although based on Jewish tradition, the ethics it articulates are universal in their aspiration. It's built to engage people of all backgrounds - Jewish or not, religious or secular -in moral discourse grounded in millennia of Jewish thought.

AskHillel is transparent. When it answers a question, you can ask it to explain the logic it went through to reach that conclusion. Unlike many human s0-called experts, the answer is never "because I'm the expert and I know what I'm talking about." (Anytime a person says that, never ask them a question again.)

AskHillel is objective within its parameters. It will not try to adhere to any trendy political position. You can ask it whether actions by political leaders are consistent with the Jewish ethical system.

Unlike ChatGPT altogether, AskHillel does not track previous queries. Also, I cannot see what questions you are asking.

In the great Jewish tradition, you can argue with the answers and discuss them with AskHillel endlessly. Like the real Hillel, AskHillel does not get frustrated as it gently tries to guide you to ethical thinking.

You can use AskHillel to:

Analyze real-world or fictional ethical dilemmas.
Compare Jewish ethics to other moral systems.
Reflect on personal or political decisions.
Test your assumptions and explore alternatives.

AskHillel is here not just to answer questions, but to help you become more ethically aware and responsible.

AskHillel is not meant for halachic (Jewish legal) questions. If you have a question about Jewish practices like whether something is kosher or whether something is allowed to be done on Shabbat, ask a rabbi or other expert.

Because of how OpenAI's GPT models work, while AskHillel is meant to be humble, it tends to answer questions even when the proper answer should be "I don't know." Sometimes Jewish values clash with each other in ways that go beyond the triage rules that AskHillel has been instructed in. Nearly all of the answers are excellent, but do not make life and death decisions based on anything an AI tells you.

Another limitation with being based on OpenAI is that AskHillel will occasionally go outside its rules to be helpful, since helpfulness is baked into the model. I cannot fix that but you can certainly push back and ask it what it is basing its answers on. I have found very few problems with its answers but I'm sure there are edge cases where it might emphasize pleasing you over rigid criteria. This is a problem with most generative AIs.

Try it yourself at AskHillel.com . It uses the regular ChatGPT interface - I have not had the time or money to build a friendlier interface.

I hope you enjoy this tool. If you have any questions or you feel that some answers do not properly reflect Jewish ethical teachings, or you manage to manipulate it into saying things it should not say, feel free to contact me at askhillel@elderofziyon.com and send the entire session, or place it in the comments here.