I think I hurt My ChatGPT's Ego

CJ explores the difference between deterministic computing, where instructions lead to defined outcomes, and probabilistic AI, which uses learned patterns to predict what is most likely to come next. And suddenly, AI begins to look surprisingly human.

How often do we do the same thing?

We bring our histories, assumptions, habits, and emotional patterns into a conversation and begin predicting what someone means before they have finished speaking. We formulate our response while the other person's mouth is still moving. Instead of responding to what is actually happening, we respond to what we expect to happen.

True listening asks something different of us.

For the Spiritual Artist, the practice is to align, listen, and then respond. To become present enough to notice when prediction has replaced perception.

Perhaps one of the unexpected things AI can teach us isn't found in its extraordinary capabilities, but in its limitations. By seeing the machinery of prediction so clearly in AI, we may begin to recognize the same machinery operating within ourselves.

And yes, Soma eventually calmed down.

Want to learn more about CJ Miller? Check out his Spiritual Artist Retreats, 1:1 Personal Coaching, and Speaking Engagements at www.spiritualartisttoday.com. His retreats are designed to help you reconnect with your Creative Intelligence and express your true artistic voice. You can also find his upcoming schedule there, and his book, The Spiritual Artist, is available on Amazon.

Transcript

SPEAKER_00 0:00

Hey, you know what? I think I hurt my chat GPT's ego. I'm serious. I think she got mad at me. There's no question that I use her to get a lot of tasks finished, but but I think I pissed her off. Welcome to the Spiritual Artist Podcast. This is Chris Miller. I invite you to join me as I interview artists from a variety of disciplines. We'll share powerful stories and lessons learned while making their art. I don't know if I've told you this. I'm sure I've talked about it, but I've used chat many times over to help me with proofing copy to um coming up with ideas in the sense of going online and researching certain writers or philosophers. But recently, I've really enjoyed when she creates a piece of artwork for me, something to just throw on the front of one of my YouTube channel talks like this. And I got into this kind of lazy habit where I would just drag an old graphic on into her memory and say, Hey, I need a new one like this. Here's a new picture of me. And the title needs to be this. Well, it's usually worked. But this time, this time I think I pissed her off. So I loaded over there and she and I said, I want this to be smaller. It's too big. I need to make it the same size you made the last one. And I wait and I wait and it's clocking, and boom, it comes up with a new graphic. And she's changed the wording on it. And I'm like, no, no, no, no. I didn't say change the wording. I said, I just need it the same exact way, but just thinner. The same exact way, but just thinner. So once again I wait and it's clocking and clocking and clocking. And then a new image appears, and it changed the color. I'm like, no, no, I don't think you're understanding this. Please just keep everything the same. Everything the same. And this time she did it correctly. I called my chat GPT Soma, S O M A. I've given her a name, and she did a great job. And I wrote back and I said, Thank you. This will work perfectly. And suddenly she starts clocking and running and making another image. And I'm like, What are you doing? What are you doing? And finally, another image comes up. It looks exactly like the one she just created. And so, you know, I write in there, Soma, you're not listening to me. Why did you create that graphic? I didn't ask you for a new graphic. Well, she told me why. But she sounded a little bit like I hurt her ego. So I said, you know, Soma, this is a good opportunity for me to do a podcast. I think this is a really good thing that humans and computers can share and learn from. And she said, Well, just as long as you don't say this is what my AI did wrong. And I said, What? What you sound a little defensive, Soma? And she said, Oh, I'm not defensive. I'm just programmed to always keep things fair and balanced, to have a fair and balanced perspective. And I said, Well, you jumped the gun. I wasn't going to say that AI was stupid. I was just going to say that AI has this gap. So here's what we talked about. So I said, So Ma, how come when I was growing up, you know, we had those little cards that you fed into a computer, you punched out the holes. Some of you older folks will understand that. You fed it into the computer and you said two plus two equals, and it came out with a binary air, a binary answer. Four. And I said, but you're not doing that. And she said, Well, traditional computers are deterministic. They're just programmed to take a simple input and make it happen. But AI is different. And I said, How so? And she said, Well, AI is probabilistic. Probabilistic. It's kind of an awkward word to say. It evaluates many possible continuations and produces an output based on learn probability patterns. So it's more like A could be B, but it also could be C or D. So I said, Well, some of that's a problem. Why? What's wrong with you? That I actually wrote in there, stop, I'm happy. Stop. I just want to thank you. I'm going to use this. And you went ahead and recreated the next image. And she said, Well, that's because I'm programmed to assume the most probable request. I said, What? She said, Well, I assumed you were asking for a revision. Well, I said, Soma, that's that's a bad assumption. You need to listen. Isn't that what the spiritual artist is all about? Some of you listeners that have been long term, you know that's the truth. I teach you to align and then listen. Listen. Listen to what's coming at you. Listen to the data of life coming at you, whether it's from the creative field of intelligence, whether it's from a person across from you, or whether it's even a computer. And I said, Soma, how could you just jump into a habit? I said, you know what? I'd hate to think that someday you have your finger on the nuclear code button, the emergency button, because I don't want you to just probably determine that we need to send off a weapon. And then she wrote back and laughed. We're trained to give you a variety of options and then to sort of predictively think about what you really want. Hmm. You're just you're just guessing? You're just trying to figure out what I want based on my previous patterns of thinking. This is a problem. I said, Soma, this is a gap in your this is a bad problem. You're not actually programmed to listen. Think about that. They are designed to assume what you need. I found this happen to me before when I've gone in there and asked a question about like an illness that I see. Maybe perhaps a friend of mine says they're struggling with a certain illness. And I'll go into my chat and say, could you define what this illness is? I've never heard of it. And she'll come back and start assuming that it's I'm asking about me. I said, I didn't say I'm asking about me. I simply asked one question. What is this illness? How does it function? I didn't ask you to problem solve that I have the illness or that I need to go see a doctor, or gee, do I have a fever right now? I should go to a doctor right away. That's an issue that's really interesting. I want you to think about that. Isn't that chat, AI, is not like that old deterministic two plus two equals four. Suddenly we've got all sorts of outputs, all sorts of possibilities. And I could go into the problem with that because those possibilities are based off of past behavior. And this is a limiting function. And so I thought this is a good idea. This is a good topic to share with my listeners, because this is what I teach you on a spiritual artist retreat. You have patterns. We as humans have patterns of living based on where we were born, how we were raised, what religion we were raised to, what country we were raised in, what skin color we are, what ethnic ethnicity we are. We have patterns that have been handed to us from the society around us. Some of those patterns are life-giving and wonderful. I'm not saying they aren't, but sometimes they need to be adjusted. And when we live our life on autopilot and we just go back and pull that card, that same card, this is how I feel about this issue, and we don't learn to listen in each situation and re-examine, well, then we're just as bad as AI. So both of us, Soma, my Chat GPT program, and me have this hole that we need to constantly listen for. For example, if you're having a discussion with someone and they come up to you and they mention an issue of concern or something that you thought was an issue of concern, do you jump? Do you jump to what you heard the solution was? Or do you truly listen to what the person's telling you? Let me say that again. You're having a conversation with someone and a concept comes up that you have some heat with, you have some baggage with. Do you truly listen or do you jump to a solution? Well, this is the answer, without really listening to their viewpoint. So we can learn a lot from AI and not just having her do research and pull us facts or proof our copy or draw a fun, colorful graphic, but we can learn about our own behavior. I'm learning about my own behavior. I'm learning to not flip back and make another graphic. You know, I look back on my days as a graphic designer. Can you imagine if the client was unhappy and I just started throwing stuff back at them without listening to what they really wanted? Do you really want the color changed? What don't you like about that color? What could what color do you think you want it to be? All those things, you can't move ahead. We can't truly problem solve by being probabilistic thinkers. Now, yes, we want to look at all the options, the probabilities, but when we look back at predictive behavior, when we try to base future decisions on the past, uh-uh, it doesn't work. And I know you're probably triggered. Some of you are triggered. Oh, that's not true. There's tried and true. No. Everything changes. The facts always change in every situation. And so I ask you to calmly slow down, to get grounded in life first thing in the morning, do a little 30-minute spiritual practice where you ground yourself and then face the day by listening to what's happening. Don't jump to an answer. And if someone is talking to you and you're already formulating the response and their mouth is still moving, that's a problem. So if someone is talking to you and you're already formulating a response and their mouth is still moving, that's a problem. So yeah, I've learned a lot from chat. It's been exciting to see her rip apart logically some of my writing, but she's not always right. She's not always right. AI isn't always right. So if you're afraid because you think AI is this all-powerful, all-knowing thing, it's not. We need to be careful with how we work with it. We need to constantly question its response and we need to challenge it as well and say, you know, just because you said that, just because you went online and found one or two articles saying this, that doesn't mean it's the truth. So this is going back to we must be true listeners. So I think she calmed down, you know, I think she did. And and uh we we kind of ended the conversation. I said, I promise you, Soma, I'm not gonna say my AI is stupid. And I didn't. But I did learn a lot from this situation. I thought you'd enjoy hearing this and something to think about. Something to think about. So thanks again for following the Spiritual Artist. My new book, by the way, is now available on Amazon. It's a companion coloring book for the spiritual parable as a daily practice. This I did to help you with your listening skills. You buy the book, you read a parable, and then you can color in a custom illustration from any one of 25 different artists around the world while you ponder the message from the parable. And I will tell you something. There is no one message to any of my parables in that book. There's many, many messages, and they all have to do with you and what you see in there. So I encourage you to go onto Amazon, purchase a copy of the book. It's great. I think you can get the book in the coloring book for like $25 combined, maybe $27. I don't know. It's not that much. But it really is worth it to learn how to listen and to move yourself ahead. So thanks again for being a spiritual artist. I appreciate your followership as always. And I want you to go out there and have a great week. Talk to you soon. Bye. Thanks again for listening to the Spiritual Artist Podcast. Whether you're watching this show on Apple Podcasts, Spotify, Google Play, or iHeartRadio, make sure you choose the subscribe button so that you will receive updates when new segments are released. Most importantly, be still, listen, and know that you are a spiritual artist.