AI’s Inward- and Outward-Facing Problems

AI’s Inward- and Outward-Facing Problems

Last June, more than 300 people showed up for church services in Germany. What made this gathering so special, and drew so many attendees, was that the service was run entirely by artificial intelligence. As a rabbi, I think, “Well, AI’s not going to come for my job” – and then I read in the paper that 300 people came to see this AI, who appeared on a big screen in front of all the parishioners in St. Paul’s Church. He was a bearded man, alongside other a few other AI females and males who accompanied him in the service, which was 40 minutes long. So also, hey, some people do it in 40 minutes – that’s even faster. (That was just a joke. I wouldn’t do this service in 40 minutes.) And it included a sermon, music and prayers.

Despite the AI’s expressionless face and monotone voice, it reportedly held the congregation’s attention. Obviously, the large audience was probably more because it was a new thing – just the strangeness of it brought out a lot of people. But this is something for us to also consider, because right now it might sound ridiculous, but maybe AI will get better. And it will get better. And they won’t sound so monotone, and it will be able to hold our attention even more. And maybe pastors and rabbis might start being replaced by artificial intelligence. So this is only the beginning of all the ways that AI can really surprise us in its ability to perform tasks.

In order to talk about AI in a Jewish context, I want to point out that there will be aspects of Judaism, such as who can give a meaningful sermon, or how ritual items can be made and still be considered Kosher, that we take for granted now, but that are going to have to be discussed and decided on in the near future. And this is what a lot of Jewish thinkers and leaders are coming to terms with – these are real considerations for the future.

So some examples of this are: it takes over a year to write a Torah, and it’s done by hand. The Sofer has to get an extensive amount of training. They do it by hand. It’s costly. And you could say it’s excruciatingly inefficient, compared to what we can print ourselves. So what if AI could do it? Maybe the Torah is too far. What about tying tzitzit? What ways might AI-created ritual items change our ideas about the concept of what’s sacred and Kosher? And who decides these matters? Who decides what’s allowed and what’s not allowed?

Another example that I’ve been thinking about, too, is that we have AI, such as the Apple Watch, that’s developed to monitor our vital signs. And if somebody has fallen, or if they’re having a heart attack, God forbid, that AI can register these levels of stress, register that you’ve fallen, and contact emergency services. So it can save a life, and it will only get better and better at it. Does that mean that it can override the prohibition of having such a monitor on during Shabbat? What if somebody lives alone? What if they’re very prone to falling and they need the help, but they observe Shabbos?

So these are the considerations that we’re having right now – let alone what will be coming in the future. And these feel like new questions, because the technology is relatively new for many of us, but the questions – which I find comforting – are philosophically the same as the ones that our ancestors dealt with each time that they were confronted with new technology. We are an old tradition, and we have been confronted with all kinds of modernization and new ideas, and have made sense of it within our own Jewish lens.

One of the thinkers that I’ve been following a lot has been David Zvi Kalman. He’s a writer and a scholar, and he offers two ways to frame the conversation around AI and Judaism. He recommends framing our questions as inward-facing questions and outward-facing questions. And that’s how we’re going to frame the conversation today – inward-facing questions and outward-facing questions.

So, starting with outward-facing, we might ask: how might Jewish ethics and values influence AI’s development and usage? Examples of this are questions that are related to how much responsibility humans have over AI and how it works, and over technological developments. So what do Jewish ethics and values teach us about what is human, or what counts as human? What if AI can pass as human? Does it have the rights of a human? How do we deal with the  output of an AI that can pass in the real world as if it was human-created? These kinds of outward-facing questions are dealing with transparency.

Also, the development of AI, which I think can be hard for many of us to wrap our minds around, is astonishingly quick. Things where we think, “This is ridiculous, this is never going to sound like a human being,” but then it’s giving a sermon in Germany to over 300 people – it will get better very quickly. So we shouldn’t assume that we know the limits of artificial intelligence. It doesn’t mean that we should necessarily be afraid, just that we shouldn’t assume that we know what the limits are. And there’s still time. That’s part of what many leaders, especially Jewish leaders in this conversation around Jewish ethics and artificial intelligence, are saying, is that there’s still time, although it’s limited, to discuss what AI should and should not be allowed to do.

As ethical consumers of AI, we have to be the ones to set limits on AI activity. How do we do that? We do that through legislation, and putting our voices out there about things that are wrong. One topical example of ways in which AI is creating problems that we should be concerned about: AI is generating fake images about the war in Gaza, and it’s very hard in many of these pictures to know right off the bat, especially for unaware viewers, that these are fake images.

It was pretty recent that the foolproof way to know if something was AI was to look at the hands. “You should look at the hands in the image.” A lot of times, the AI was having a hard time creating images of people’s hands. A lot of times they looked mashed-up or they had six or seven fingers. It was really scary stuff. And it was almost a foolproof way to find out whether or not an image was real.

That’s not really a problem anymore. They’ve really been able to fix that glitch. And so we have to have other rules around images on the Internet. This is like the Wild West, right? It’s hard to control, and it’s hard to figure out what’s real and what’s not real. But it’s affecting people’s lives.

Another way it’s affecting lives that we need to be aware of: recently I learned that there are fake nudes of real people. This is a problem that’s happening in some high schools. Young girls are having their face taken and put on a nude. And it’s very difficult to differentiate that it’s not their body. So there has to be some legislation that’s calling this out as child pornography, that’s protecting people and our children from some of these really horrendous and just horrible things that can happen.

So there are a lot of ways to go into these outward-facing questions. And Judaism has great ethics and values to teach us, right? “You don’t put a stumbling block in front of a blind person” is one way to say that you shouldn’t have customers sign these long consent forms that nobody’s reading. I’m not reading them. I’m going as fast as I can so I can just get into the website. But in a way, couldn’t that be seen as putting a stumbling block in front of a blind person, tripping somebody and causing them harm, because there’s no possible way that they’re reading all of this information? So there are a lot of ways in which Jewish thought can influence AI.

What I find more interesting for our conversation this morning are the inward-facing questions. Now, of course, I find this interesting because this is where it’s more complicated, I think, for us. I think it’s easier for us to say, “These are the Jewish ethics and values that can inform us how we should be acting in the world.” That feels, to me, easy. The harder question, to me, is the inward-facing question, of how might artificial intelligence influence Jewish life and thought.

So some examples of that are: can one outsource the fulfillment of a mitzvah to an AI? Can AI be used to circumvent Shabbat regulations? Can artificial intelligence be a decider for halacha, for Jewish law? And as I said earlier, these feel like new questions and new problems. But our tradition is so old, it has dealt with a lot of things in the past that have had to be confronted and dealt with.

So one example, which also was written by David Zvi Kalman – this article for the Jewish Exponent, where he talked about how matzah production by machines was seen as a large threat. Or, I shouldn’t say “threat” – it was concerning to many leaders in the Jewish community. So over 100 years ago, there were dozens of European rabbis engaged in the debate over whether or not machine-processed matzah, which was a new thing for them, was considered kosher. Up until this time, a person who knew the rules, knew Torah, had to make matzah by hand, hopefully with the intention of creating something that is used in a ritualistic, holy way.

But just as we are finding with artificial intelligence, the machines that can make matzah promised huge savings – savings of time and money. But as I said, the leaders of the Jewish community were very concerned, because they had not used machines for this before. It was a new thing, and how did they know that if it was going to work?

In 1859, a collection of letters was compiled from prestigious rabbis in Europe alerting the Jewish people that matzah made with a machine should be considered not kosher for Passover, that it was “no better than a loaf of bread.” For us today, this feels a little quaint – because how many of us have ever been to a Passover seder where there wasn’t a box of Manischewitz matzah, or another store-bought brand, on the table? It’s very common and understood to be Kosher, and accepted in most homes.

But we understand this now because of the distance of time. This new machine technology in the 19th century was scary and concerning for many people. What would the implications be? The rabbis wanted to know: are the machines as reliable as the people? We trust people – how can we trust that the machine will do the right job?

The other concern was: what if bits of dough got stuck in the gears of this machine and just sat in there? Leavening could be happening for hours, and the factory could already be running another batch, and without even realizing it, something closer to bread could be shoved into this new batch of matzah, and nobody would know the difference. So these were real concerns. What if the trays in the machine warmed the dough too fast? Without proper oversight, the rabbis argued, “How can you really trust your own food?”

Some also objected to the loss of physical human touch. Jewish law states that matzah is supposed to be made by people who know how to bake matzah. And it’s not that difficult of a process – it’s just a few ingredients and it’s 18 minutes. But it was concerning – if a machine is not sophisticated and doesn’t know anything, how can you trust it?

But there’s the argument that maybe machines, over time, can be more reliable. And then there was a response to the rabbis’ letters in the 1800s that didn’t take very much time at all. They wanted to cancel the alert to the Jewish people then, to say it’s okay to have matzah that’s made in a machine. Some argued it was possibly even better, because the machine wasn’t going to make a mistake. What if a human being making matzah lost track of time, or they didn’t do something correctly? With a machine, our understanding of how it is supposed to work is: it gets it done right, every time. It’s not going to start thinking about its day and daydreaming and losing track of time. It’s also less expensive for everyone and makes its product more widely available.

So here’s what Kalman writes, which I thought was interesting and I wanted to read to you:

“Making matzah locally could have been a way to feel connected to the ancient Israelites, who left Egypt so fast that they didn’t have time to make anything else. Instead of emulating this ad-hoc food, we optimized it for cost and efficiency, in the process turning matzah into just another specialty cracker on the grocery store shelf. Was it really worth it?”

He argues, further:

“It’s probably a bit much to say that OpenAI is just a modern Manischewitz, but the parallels between the debate about machine-generated matzah and the present debate about machine-generated everything are useful for considering how short-term policy choices around AI won’t necessarily capture all of the technology’s long-term effects on how human beings want to spend their time. When we relinquish an activity to an AI for economic reasons, we may eventually come to believe that humans are no longer qualified to do the task at all.”

He’s a very pro-machine, pro-AI writer, but that’s a bit of his warning, saying that with some things, when we estrange ourselves from the process, we lose that bit of spiritual experience of creating them ourselves.

So, in conclusion, with these two large ideas, these inward-facing questions and these outward-facing questions, we should note that Judaism modernizes constantly. In some movements, it moves very fast. In other movements, it moves much slower. But Judaism is always adapting to what’s being put in front of it. And oftentimes Jews are involved in creating the new science and technology that’s being made. It’s constantly evolving. So we have questions like this that are going to become more and more urgent to answer. And some of them are small, such as whether Kosher meat could be slaughtered by a machine or an AI – and what if it’s proven that this is more effective, and we find out that the animal suffers even less, and that would that make it more Kosher than if a human being slaughtered it?

And these are the kinds of questions that are going to come up, and there are going to be different answers for them. It gets to be about how we, as Jews, are in conversation, as well, about the merits and the problems of artificial intelligence. Not to mention, too, the quality of life for people whose jobs are on the line, for many of these types of factory jobs – we have to consider, too, what happens to people who are possibly going to be left behind in this development that’s happening at a breakneck speed. So I hope that this has put some more questions in your mind. I know there are not really that many answers, right, because it’s always changing. But the questions are really important. Especially as Jews, we’re comfortable with questions.

And so as we go into lunch, I invite you to talk with one another, and to talk about this one question. What do you think is the short-term and the long-term interaction of Judaism and technology?

Shabbat Shalom.

(This post is part of Sinai and Synapses’ project Scientists in Synagogues, a grass-roots program to offer Jews opportunities to explore the most interesting and pressing questions surrounding Judaism and science. Jeanine Jankovitz is Rabbi at Congregation Beth El-Ner Tamid in Broomall, PA. This was adapted from a Shabbat sermon given at that congregation on November 4).

0 Comments

Add a Comment

Your email address will not be published. Required fields are marked *