The rabbis of old may not have had to grapple with AI but they had to face innovation in their time. As we look to navigate the challenges of advanced technology, our sources may help to point us in the right direction.
Some of those challenges were set out in the 20th Louis Jacobs Memorial Lecture which was recently delivered by the new senior rabbi of New North London Synagogue, Daniel Nevins, who took up his post earlier this month. He has addressed some of the emerging issues as a member of the Jewish law and standards committee of the Conservative movement’s Rabbinical Assembly in the USA, writing a a chapter on AI in his 2024 book Torah & Technology.
Our ethics depend on the idea that we take responsibility for our actions. But the arena is becoming more complex when we start to rely on machines that can act autonomously.
In warfare, militaries are increasingly using AI to select targets. But if a machine wrongly identifies a civilian site, who ultimately is to be held to account?
Drones used to be controlled by drone pilots, Rabbi Nevins explained.
“Pilots would see the video feed of what the target was doing and they would either authorise the attack or call it off. But then each side learnt to jam the communications of the drone so they couldn’t be flown that way. So they figured another way to do it which was to equip the drones with their own mapping software, their own targeting information and to set them free autonomously to go and attack the target.
Unmanned commercial drones carrying air bombs in a battlefield. Illustration of the concept of using a swarm of drones as weapons in warsGetty Images
“But what if the target, say an armoured troop carrier in a field by itself, now has moved and it is suddenly in the middle of a town surrounded by children. Will the drone operated by AI know to call off the attack and not to blow up that target where there are children right nearby?”
Or take another scenario. The first generation of driverless cars would meticulously observe speed limits. “Then they learnt they could match the average speed of the other cars. So let’s say they are speeding and they are pulled over by the police, who pays the ticket? The user is asleep in the back seat. They have got a pretty good defence. It wasn’t me, it was the machine – go after the engineers who created this.”
And what would happen if the speeding car was involved in an accident?
In the new field of “machine ethics”, AI creators are trying to integrate ethics into machine decision-making. But how do we agree the values that are to be incorporated? There needs to be “guardrails in the system that prevent you from just following the majority,” Rabbi Nevins argued. To take the driverless car, for example, if the majority of drivers on a particular road are breaking the speed limit, it might calculate that this would be the average speed to maintain.
“I feel any time we have got a human life at stake a human needs to be in the mix because humans feel responsible and humans can be held responsible for their errors,” he argued.
There were areas of halachic analysis, he suggested, that could be helpful in formulating an approach, citing in particular nezikin, the laws on damages, and shlichut, agency – when someone acts on your behalf. When the rabbis looked at the instance of an ox that strays into a field and gores another animal, they differentiated between an ox with no previous record of rogue behaviour and one that had acted harmfully before in determining the level of liability.
So there might be difference between a machine the user could reasonably expect to operate safely and another where problems had previously arisen.
A Waymo robotaxi in California (photo: Getty Images)Getty Images
But new technologies also raise other questions, too. On a lighter note, he imagined the possibility of an AI-controlled coffee-maker able to recognised the people in a household and when each one came into the kitchen in the morning, immediately greeted them with their favourite coffee. Even if it were to be halachically permissible on Shabbat, would it detract from the special nature of the day?
“What is lost when you start to have a Shabbat environment where everything that you want can be anticipated or provided in a way that is not strictly forbidden?” he wondered.
Rather, he argued, in this new machine age Shabbat and festivals could become more significant in reminding us of who we are. When algorithms complete sentences in the messages we send, recommend what TV shows to watch or books to buy, “sometimes I worry we’re forgetting to live our own lives,” he said.
What was distinctive about human beings lay not in any skills that we possess but in our relationships, he argued, “sitting in a room with people, with family and friends, expressing concern for one another, knowing that all of us have a limited term on earth… knowing we have an obligation to one another, knowing that we can feel pride for good things we have done and shame for mistakes that we have made, knowing joy and knowing pain and suffering, these embodied experiences of human intelligence are what I think makes us distinct from the machines.
“Which means if we want to preserve human distinctiveness, we must really concentrate on human relationships.”
Shabbat is a day that liberates us from thraldom to phones and screens, and enables us to cultivate relationships. When you can go for a walk without a phone in your pocket or have a conversation with someone without the prompts of assistive technology, he suggested, “you will find yourself more free and perhaps more joyful”.
Rabbi Nevins’s lecture can be accessed here
Torah & Technology: Circuits, Cells & the Sacred Path is available from Izzun books
To get more from judaism, click here to sign up for our free daily newsletter.
