Leap of faith: Approaching AI from the Jewish perspective

In a recent interview, the Israeli author and historian Yuval Noah Harari said that we are the first generation to have no idea what the future will look like for our children. Artificial intelligence (AI) is one of the most urgent factors.

One way of approaching AI from the Jewish perspective is by re-reading texts about golem, an anthropomorphic being in Jewish folklore created entirely from clay or mud. The most famous golem narrative involves Judah Loew ben Bezalel, the late 16th-century rabbi of Prague. In this story, a golem was created as a helper and protector, but due to unforeseen development, the creature went out of control and became violent. Luckily, the rabbi embedded an ‘off switch’ in his design and was able to stop the dangerous monster.

Perhaps the first lesson from this story is having the ability to turn the AI off in case of a serious malfunction.

Rabbi Tzvi Hirsch Ashkenazi (Chacham Tzvi) also offers a provoking commentary. He refers to a rabbinic teaching: “The works of the righteous are their offspring.” This suggests that the creation of a righteous Jew could be considered Jewish (Responsum Chacham Tzvi 93). In other words, the robot can be Jewish if the robot’s developers (read ‘parents’) are Jewish. It could be a funny joke if it weren’t so relevant.

AI models use deep learning techniques and massive datasets of human knowledge to process and predict language similarly to how a human communicated before. Therefore, if you limit the AI model to only the Jewish knowledge dataset, it might be considered Jewish.

In her book Golem, Maya Barzilai argues that golems do not share the most fundamental part of the human condition – they are not born and do not die. Likewise, robots might be human-like and know more than any rabbi, but they can never fully understand humans. How can someone who never dies read Mourners Kaddish?

As a Progressive rabbi, I welcome AI with its promise of helpful assistance. However, this powerful tool must be used with a great sense of responsibility. AI-powered machines can be invaluable in many areas of our lives, but I don’t think they can be meaningful in such a personal and intimate space as prayer when we humans speak to the deepest part of ourselves that we call God.

We may not have had golems in our reality, but we do have software corporations and countries like China working incredibly hard to create the most powerful AI-powered tools and weapons today. While we may not be able to stop or pause AI development, we must be prepared to face the era when human-like robots will be around. I don’t have all the answers, but I know that we must engage with urgent questions that AI forces us to answer: what new opportunities and limits will AI bring to humanity, what boundaries can we put in place to mitigate potential threats and who should decide on what these boundaries are?