Shana tovah and Shabbat Shalom! I’m so grateful to be here with you to celebrate this New Year.
Occasionally, some well-meaning person will ask me if I write all of my own sermons. The people asking this question have either a) never interacted with me when I am trying to finish a sermon or b) imagined some giant database from which all clergy download their content.
Once a conservative colleague offered to go halfsies with me, each of us writing two of the four High Holy Day sermons, but I knew it wouldn’t work. My sermons are deeply personal to me, and I try my best to make them deeply personal to this community as well. Writing is one of the parts of my work I that take the most pride in, and expend the most time and energy on.
But towards the end of this past school year, having just finished a huge project writing about the weekly Torah portion for the URJ (so that OTHER rabbis could download them!), I felt like I was out of ideas. So I did something I never imagined that I would do: I asked ChatGPT to write a Shabbat sermon for me.
ChatGPT is a natural language processing tool driven by Artificial Intelligence. I am just beginning to understand how it works, but given certain instructions, ChatGPT can use predictive technology to generate writing such as emails, essays, or code. You might already have seen earlier forms of this technology in the “autocomplete” functions in your email or text messages. The generated text can then be refined with further instructions, resulting in writing that sounds… almost…human.
I had no intention of ever using ChatGPT. But curiosity—and end-of-year procrastination—got the best of me, and I typed in “Write a sermon for a Reform Jewish congregation about Parshat Balak.”
That’s all I asked. And that’s what it did. In under a minute. It was actually more concise than I usually am—making its point in a mere 500 words. I was simultaneously delighted and terrified.
This exercise triggered in me an existential crisis and an ethical dilemma. The former taking the shape of: Was I about to be replaced by a machine? And the latter taking the shape of: So….is it okay to use this tonight?
I wished, as I often do, that I could call my dad. He would have been fascinated by these questions, as well as by the new technology that precipitated them.
But considering it was already Friday morning and there is still no technology for communicating with our loved ones in the afterlife, I had to figure it out for myself. I decided that I could deliver the sermon only if I disclosed its origins first, so that we could discuss it afterwards. Some of you were there that evening and might remember.
The sermon opened with “Shabbat Shalom, my dear congregation,” and everyone chuckled. But as I continued, the laughter abruptly stopped. Apart from some heavy-handedness with the phrase “as Reform Jews” (a nod to the original prompt), what ChatGPT wrote was not so different from what one might expect to hear at any synagogue on a Friday night. The sermon was, ironically, about the power of language. This was its closing paragraph:
“As we engage with the teachings of Parshat Balak, let us commit ourselves to opening our eyes and hearts to the beauty and worth of every person we encounter. Let us strive to see beyond appearances, to challenge stereotypes, and to embrace the diversity that enriches our lives. And above all, let us be mindful of the power of our words, using them to build bridges and promote understanding in our communities and beyond.”
There was a charged silence when I finished speaking, after which a few people quickly jumped in to reassure me that my job was safe.
And this led us in a direction that I’m not sure even ChatGPT could have predicted—a conversation about what it means to be human, and why our lives, and particularly our spiritual lives, still require a human touch.
There are many discussions and debates to be had about the implications of such technology for our lives, our work, and our society. But tonight, let’s leave that to the journalists, politicians, computer scientists, and ethicists.
As we prepare to enter a new year as a spiritual community, let’s take a moment to consider this question: What makes you, you? What is it about you, or the person sitting next to you, that cannot be replicated or replaced by a machine?
For the community that was gathered that Shabbat, the answers quickly crystallized into three categories: the wisdom we gain from our experiences, our creativity and sense of humor, and our capacity to build relationships with each other.
Let’s start with the wisdom of our experiences. The very first lecture I attended in rabbinical school was titled, “You don’t need to know it, you just need to know where to look it up.” At the time, my professor, Moshe Silberschein, was talking about dictionaries, encyclopedias, and leather-bound rabbinic volumes with gold lettering that had yet to be digitized. Maybe a database or two available on CD-ROM. It was 2003, after all.
I took this advice to heart. When my dad would ask me a rabbinical question, I’d tell him I found the answer with the help of “Rabbi Google.” With mock anger, he’d say, “Well, I could have done that myself.” And I’d respond, “No you couldn’t, because you wouldn’t have known what to Google.”
Knowing how to look things up requires its own kind of knowledge and critical thinking. We need to know enough to ask the right questions, to seek out the answers in the right places, and to know whether and when we’ve found what we’re looking for.
The same is true of Artificial Intelligence. Not only is it taught based on content that we’ve created. It still needs us to ask the right questions and to know what a good answer looks like.
There is also a kind of learning that the machines can’t do. They may be able to imitate our voices, but they cannot gain the wisdom that each of us derives from our own experiences.
Artificial Intelligence learns by consuming content, looking for patterns, and learning rules. And one of the most important things I’ve learned in my years as a rabbi is that you cannot simply take every person or situation, plug in a formula, and produce the right response every time.
The Irish rugby player Brian O’Driscoll once said, “Knowledge is knowing that a tomato is a fruit. Wisdom is knowing not to put it in a fruit salad.”
Part of human wisdom is learning that every situation might be different from the one that came before it. Not everything fits neatly into a pattern, or follows a hard-and-fast rule. We need to listen, to pay attention, to understand the nuance of each situation, even as we look to our collective wisdom for guidance. And it might take a few tries to get it right.
As Mark Twain once wrote: “Good decisions come from experience…Experience comes from making bad decisions.”
Part of being human is recognizing that sometimes we get it wrong, and we have to try again. This is the central idea of this season. Teshuva, or turning, is our way of putting ourselves back on track after we mess up. It is a moment when we ask ourselves: What have I learned from my experiences this year? What patterns have I noticed that I don’t want to keep replicating? How can I be compassionate to myself so that I can move forward? And how might what I’ve learned help me to be more empathetic to the people around me?
Before things get too serious, let’s move on to our creativity and sense of humor. When I’m writing a sermon, I like to sneak in a joke right about here, because, if you laugh, I know that you are still paying attention.
One of my favorite movies growing up was the 1986 comedy Short Circuit, in which a prototype for a new military robot is struck by lightning and becomes convinced that it (he?) is now “alive.” He sets off on an adventure to acquire new “input,” and to avoid being “disassembled,” which he equates with death. We watched this movie so many times as a family that its dialogue became part of our household vocabulary.
Throughout the movie, the programmer tries to catch up with the robot and prove that he is, in fact, just a robot. Finally, the programmer decides that the only way to settle this once and for all is to tell the robot a joke. He chooses, of course, a joke featuring a priest, a minister, and a rabbi. It’s not very funny (and looking back, it might have relied on some problematic Jewish stereotypes), but that’s beside the point. The robot…laughs. This laughter, which the programmer describes as a “spontaneous emotional response,” is proof enough that the robot is, in fact, alive.
I wonder what might have happened, however, if the programmer had asked the robot to tell a joke. Because ChatGPT…is not very good at this. There are programmers around the world feeding content to various AI systems so that they can learn the patterns of different genres of humor and how to replicate them. This is, by the way, also how most humans learn how to be funny, hence both kindergartners and computers usually start with Knock Knock jokes.
When I asked ChatGPT to tell me a knock, knock joke, it had one at the ready. A “classic” that we’ve all probably heard before:
“Knock, knock. Who’s there? Lettuce. Lettuce who? Lettuce in, it’s cold out here!”
But when I asked it to create a knock knock joke featuring a llama—for my niece, of course—it was able to replicate the formula, and even generate some wordplay, but it wasn’t quite able to fulfill the mission:
“Knock, knock. Who’s there? Llama. Llama who? Llama tell you a secret, but you have to promise not to spit the beans!”
Because AI is good at learning patterns and following rules, it probably won’t be long before it can generate a decent joke, the plot of a basic rom-com, or an episode of Law and Order. But as prolific TV writer Michael Schur explains, the question that underlies all artistic pursuits is: “Can you reach through the screen and grab someone by the lapels of their jacket and shake them a little bit and make them see the world differently or make them understand themselves differently?”
This question is not only the province of humor. Humor is simply a stand-in for that which we can only create from a combination of our own experiences, our ability to connect with people, and our understanding of the human experience as a whole. This is something we are seeking whenever we engage with art. And it is also something we seek in our spiritual lives. Not only do the words of our tradition hold the power to reach out of the page and across the centuries to shake us a little bit, they also have a way of shaking us up in new ways every time, depending on who is listening, and who is speaking.
In her book, Leaving Church: A Memoir of Faith, Dr. Barbara Brown Taylor writes of communal prayer:
“The words sounded different when Kline read them than they did when Kathy read them. They sounded different from the mouth of a young mother than they did from the mouth of a widow. This is because the words did not come straight off the page. They percolated up through the silt and gravel of people’s real lives, so that the meaning in them was fluid, not fixed” (Taylor 93).
And this leads us to possibly the most important aspect of what makes us human: our ability to form relationships and build community with each other. This is where I think the widest gap exists between what we require as human beings and what technology can provide. And ChatGPT agrees, saying:
“I am not capable of forming personal relationships or emotional connections because I am just a computer program designed to provide information and assist with tasks to the best of my abilities. …. My responses are generated based on patterns in the text data I’ve been trained on, and I do not possess the ability to engage in meaningful relationships.” Phew, that’s a relief!
Over the past few years, we have seen how technology can provide tools for better human connection. But we’ve also learned how much we still need to connect with other people in order to survive.
This idea goes all the way back to the wisdom teachings of the 2nd century, where Rabbi Hananiah ben Teriadon said, “If two sit together and there are words of Torah spoken between them, then the Shekhinah, the Divine Presence, abides among them” (Pirke Avot 3:2).
A few years ago, as we navigated those first online-or-socially-distant life events, I realized how much of my own work relies on non-verbal communication. This surprised me, since I spend so much of my time talking. But I felt cut off at the knees when I couldn’t put a hand on the shoulder of a student to let them know they’re doing okay and to keep breathing, when I couldn’t take someone’s hands in mine or offer them a hug in a time when words are inadequate. When someone I love is hurting, I want to bring over food, even when there isn’t a physical need for it. Or sit quietly with them, just so they know they aren’t alone. Some of this can be achieved through technology, but not by technology alone.
One of the most powerful statements of theology I’ve ever encountered came not from an ancient text or a modern philosopher, but from a teenaged student at Leo Baeck High School in Haifa, describing their worship experience: “I don’t know if I believe in God, but I sat next to you, and it felt good.”
This is why, no matter how many population studies or think-pieces emerge about how the work of faith communities is becoming irrelevant, I still believe we have something to offer to the modern world. A place to be human. A place to sit next to each other, silently or singing. A place to be seen and known and understood as the unique human beings that we are.
At this point, I thought it would be funny if I told you that this whole sermon had been written by ChatGPT. Don’t worry, it wasn’t (I’m just trying to see if you are still paying attention). But I did ask it what it might say if it were in my place. Here’s how it would have ended the sermon:
“As we stand on the precipice of a new year, let us remember that AI, like any other tool, is a reflection of the choices we make. Let us choose wisely, guided by our Jewish values, and use AI to build a more just, compassionate, and prosperous world. Shanah Tovah, may this year be filled with blessings, growth, and the wise use of technology in service of humanity.”
Not bad, right. But still, not necessarily what I wanted to say.
Soon my words will be online and become fodder for the machines that are learning how to write sermons. And it’s likely that dozens of my colleagues around the world are preaching on this very topic at this very moment, with or without the assistance of ChatGPT. But I still like to think that this sermon is one that only I could give.
Similarly, I like to think there is work to be done in each of our lives that only we can do. Right now, in your life, there is a conversation that only you can have. Someone who needs to hear what you have to say, and someone who needs you to really listen to them. Right now, there is an act of kindness that only you can do. A piece of wisdom that only you can share, from the experiences that have shaped you. Right now there is piece of art that only you can create, a joke that only you can tell. A battle that only you can fight. A gift that only you can give.
And so, tonight, we ask ourselves: What is the unique contribution to the world that only we can make? What might we choose to do this year, that only we can do?
This year, may we each find ways to bring blessings to our world, using the best of all that makes us human. May each of us find opportunities to bring our knowledge and wisdom, our creativity and humor, and most importantly, our compassion and empathy, to this new year. May each of us discover our own way to make our world a better place to be human.
Shana Tovah and Shabbat Shalom!
Rabbi Leah R. Berkowitz