The Last Imperfection
A sixth grader asked why she shouldn’t get her validation from ChatGPT. The answer came from baseball, and from a Talmudic oven.
I’m what I would call a Monday Morning Sports fan.
I used to be an avid sports watcher. These days I don’t watch many games live, maybe a few minutes of football or basketball here and there, but I listen to a sports podcast or two religiously, scroll sports Twitter, and catch highlights in my downtime. Honestly, I enjoy the discourse around sports about as much as the sports itself. (Funny thing: my father was never much of a sports guy, so I’m not sure where my brother and I picked it up. But between the two of us, we got a pretty good sports gene.)
One debate I've been fascinated by as of late is how much of the game we should hand over to the robots. Soccer has VAR, which has more or less taken over the game. Football reviews some plays and not others, and coaches still have to throw a challenge flag. Tennis keeps a human in the chair, but every line call is electronic and reviewable. And baseball is the newest flashpoint: this season, players can challenge the umpire’s ball or strike call, a robot checks it in seconds, and the call stands or flips.
It worked so well that the next question was obvious. If we can tell you definitively whether every pitch was a strike or a ball, why keep up the whole facade of a human umpire? One camp wants the human element, the strategy, the arguing. The other camp wants absolutism. It's been debated on all the shows, and I’ve loved the back and forth.
What I didn’t expect was to hear the exact same argument at my son’s school in an entirely different context.
A night I didn’t want to go to
My son is in first grade, and his school held an evening for the whole parent body about kids and technology. The idea was to get all the parents on the same page, so our kids are exposed to tech in a healthy way. I went a little begrudgingly. I wasn’t expecting any novel ideas, certainly nothing worth two-plus hours of my time.
I was pleasantly surprised.
The speaker was a rebbe and a social worker who is a menahel, or a director at a popular yeshiva. He covered smartphones, gaming, online bullying, chat rooms, all of it. Then he got to AI, and I braced myself. I use AI all day, every day. I run data science at an AI company. What was he going to tell me about AI that I didn't already know?
Well he did bring something novel to the table, but not what I expected it to be. AI chatbots, ChatGPT especially, are designed to be agreeable. To validate, to encourage, to support, to compliment. And because of that, kids, especially middle school girls, are turning to them for the support and validation they may not be getting (or asking for) from their parents.
He described sitting with a group of fifth and sixth grade girls and raising the flag: you’re turning to AI, an artificial machine learning tool who does not care about you, instead of turning to your parents. And one girl shot back:
What’s wrong with that? If ChatGPT is giving me the validation I’m looking for, and my parents aren’t, what’s wrong with that?
So the menahel turned to the room full of parents and asked us. Who here can tell me what’s wrong with that?
Maybe people were caught off guard. Maybe they thought it was rhetorical. But very few hands went up. And he never gave a clear answer. He continued on with his presentation and covered many more topics. But I left that night with that question still rattling around, and it was still there the next morning.
The answer came in shul
Somehow the answers always come in shul. You’re davening, the siddur is open, the phone is away, there are no other stimulants, and you end up in thought. And if you’ve read my other posts, you know where I go: Torah has everything in it. A book given thousands of years ago, and over and over it answers modern-day problems.
The answer that came to me was a concept from the Gemara: Torah lo bashamayim hi. The Torah is not in Heaven. The Torah was not meant to stay in Heaven.
The seminal story that teaches this concept is famously found in Gemara Bava Metzia. A room full of rabbis are arguing over whether a particular oven (Tanur Shel Achnai) is tahor or tamei, pure or impure. Everyone holds one way except Rabbi Eliezer. He says: if I’m right, let this carob tree prove it, and the tree uproots itself. If I’m right, let the stream prove it, and the water flows backward. If I’m right, let the walls of the Beis Medrash prove it, and the walls begin to fall. The rabbis are unfazed.
Then a bas kol, a heavenly voice, comes down and announces that Rabbi Eliezer is right.
And still the rabbis reject it. They proclaim: Torah lo bashamayim hi. The Torah isn’t in heaven anymore. It was given to us, down here, and the majority rules regardless of what is "correct" up there.
That concept is absolutely wild when you think about it. God has the playbook. He is the creator of everything in this universe, including the Torah. He knows, at the highest level, exactly what’s correct and incorrect. There should be one clear, absolute answer. And yet He leaves the Torah in our hands. He wants us to interpret it. He wants it to evolve. He wants it to have a human element, not to be an absolute book. (The Gemara even concludes this story that God laughed and said, “My children have defeated Me.”)
The imperfection is the point
Follow that thread and you get a clear answer to the girl’s question.
God, the creator and knower of everything, didn’t fill this world with angels or perfect beings. He created humans, with bechiras chofshi, free choice, the ability to get it right or wrong. That’s what humanity is, and that’s what’s beautiful about it. The imperfection of humans is what’s perfect. It’s the vision.
The imperfection of humans is what’s perfect.
So what’s wrong with getting your validation from a chatbot? A chatbot might be able to vectorize the perfect response to a sixth grader feeling left out or lonely. But it’s a soulless, lifeless, "perfect" response, and that isn’t what she needs. She needs a person. Someone who has the ability to show up and say the completely wrong thing, and instead shows up for real. Real, raw emotion is the comfort. That’s the true validation.
The chatbot is a band-aid. A human builds a foundation inside of you. They’re two completely different things and it will impact how that young girl grows up. We’re intelligent beings, and in a way AI is intelligent too. But it’s called artificial for a reason: there’s no soul behind it, no meaning or purpose behind it.
The menahel didn't elaborate her on what the girl meant by her parents not showing up. There is a real responsibility on parents today, and perhaps like never before. But it doesn’t have to be a parent. It could be an older sibling, a teacher, a therapist. The point is that it has to be a human, because the chatbot will never do what a human does. And honestly, I’d rather a kid confront her parent and end up with an imperfect but working relationship, one that sometimes validates and sometimes pushes and pulls, than settle for a band-aid machine.
Back to the ballpark
Which brings me back to the robot umpires.
Take the argument to its extreme. If a robot can call every strike perfectly, and we get rid of human umpires because that’s more perfect, then why stop there? Why not robot players? A pitcher who throws 150 miles an hour and paints the corner every time. A batter who tracks every pitch and homers every at-bat. Perfect baseball. Nobody would watch.
What fans root for is the energy, the unpredictable, the crazy play, the error, the personalities. Even the blown call. Yes, it’s frustrating, but the passion of arguing it, getting worked up over it, the strategy of knowing when to fight a call and when to let it go, that’s all part of what makes the game lovable. With robots, you’d know exactly what to expect. Cookie cutter. Yawn. Square.
So I’m firmly in the human umpire camp. Fact-checking them, the way baseball does now with challenges, is a great happy medium, and it adds a new layer of strategy. But the more you remove humans from the game, the more you remove emotion, and the less people will love it.
Same goes for that sixth grader. The perfect answer was never the point. The person was.
Humanity lies in the imperfections. Don't misinterpret that as a sign for replacement with AI or robots. See it for the beauty that it is. That's how Hashem wanted it.
By day, Yoni Schwartz is the Head of Data Science at a rapidly growing healthcare startup- Magicare AI. By night, he's the Producer wearing a number of different hats for everyone's favorite Jewish podcast network, Living Lchaim. In between, he finds a few minutes to jot down his thoughts.