[music]

[00:00:00] Francie Black: From my perspective is I want homeschooler kids to understand tech, I want them to understand AI. No matter what industry you go into, you don’t have to go into a tech industry, but it’s going to change nearly every industry. I feel if students can embrace that, understand it, learn how to work with it, and be at the table where it’s implemented, I want them to have that objective voice at the table.

[music]

[00:00:30] Gretchen Roe: Good afternoon, everyone. This is Gretchen Roe for The Demme Learning Show, and I am so excited to welcome back Francie Black. We have become conference attendee buddies, and we don’t live that far from each other, but Francie has wonderful information. She is always a wealth of information. Today I’ve invited her here to talk about AI in homeschooling, and I’m just delighted to have her.

[00:00:58] Francie: Thanks, Gretchen. Great to see you again, as always, and as you said, we are, I’ll call us booth buddies. [laughs] We’re booth buddies, and Gretchen and I live about what? Maybe 20 to 30 minutes away from each other as well, but anyhow, Gretchen’s wonderful. Thanks for having me. Just a quick background on myself. I come from a tech environment. I managed software development for a large Fortune 500 company in Los Angeles for many years, did worldwide software development for them, and then relocated to the East Coast due to my husband’s work, and went more into the consulting realm.

I have four children. When my kids were young, we had started– because I had a all STEM background, we started summer camps, we called them home hacker camps. We were doing projects with our kids in our basement, and some other kids found out about it, so we invited neighborhood kids, and then the newspaper wrote about our– we called them home hacker camps, that’s where we got our name. Newspaper wrote about us, and we said, if other kids outside of the area are interested in this type of STEM curriculum, we would be happy to open it up to other students.

We did get a lot of emails, and then the next year we hosted a big camp in our basement, and then after that, we moved to schools and so forth. We had a lot of homeschooled kids attending, so we ended up converting things into full semester courses so that students, homeschooled students specifically, could get credit for it. They were having the same issue as my kids, is that they weren’t getting the STEM knowledge that they needed, and I wanted to make sure my kids had that.

I did a lot of consulting work. I was on TV for many years, I was the tech time host for ABC WLOS News here in Asheville, North Carolina, live, I would go on weekly live for about six years, and do a feature tech segment for them, and then I also syndicated those out to about 100 TV stations across the country for many years. That stopped over COVID, and I haven’t gone back yet. That’s a little bit about me, but I am a pure tech, I love it, I love the innovation, I love how it moves people forward, I love how it makes some things easier, some things harder, but that’s what we do.

As a result, my kids are headed in STEM careers. Fast forward quite a few years, and we’ve got some– One thing I do want to mention regarding my kids is my one son, who I actually started the camps with, when he was in middle school, he is now with SpaceX. He’s a rocket engineer, working on the rocket going to Mars. He’s going to do a free talk for homeschooled students. Gretchen, I wanted to invite your listeners to that talk. It will be this fall. He’s going to talk a little bit about working at SpaceX, but he’s going to talk about the project that– He works on the very large rocket. It’s called the Starship. He’s a Starship engineer. He’s going to talk about the project, trying to get that rocket to Mars, which is what the goal is.

He’s also going to talk about the computer science and engineering track. If you’ve got middle, high school students that are thinking about STEM careers, he’s going to talk a little bit about that, and he’s going to give a mini aerospace engineering lesson. It’ll be a free one hour talk to any homeschooled students. We are going to cap it out at 100. If you’re interested in that, you can contact me. Probably the easiest way is to go to our website and go to the contact page. In the subject, you can put there that you’d be interested in that talk. That’s homehackercamp.com, not hackers, a lot of people put an S there, but homehackercamp.com. Then just click contact and you can submit for that. We’ll send you out an email invite when that happens in October.

[00:04:52] Gretchen: That sounds exciting. I’ll make sure that that information gets included in our show notes so that families who watch this or listen at a later point in time will have a place to go and find a link. That would be amazing.

[00:05:04] Francie: Perfect.

[00:05:04] Gretchen: That sounds pretty exciting.

[00:05:06] Francie: Yes, it’s really exciting. It’s a very unique opportunity. Of course I’m biased, but it’s fun. It’s cool.

[00:05:11] Gretchen: Absolutely. Francie, we were talking just earlier today how much the world has changed with the advent of AI. I’m a little suspicious, but there are things that AI has already made a little bit easier in my life. I’m not willing to throw the baby out with the bathwater, so to speak. Tell us what you see and where a parent who’s trying to get themselves comfortable with this should begin.

[00:05:40] Francie: AI is rapidly changing the way that things are being done. We’re seeing a little bit of it with the chat GPT, and we’re going to talk heavily about that. We’re going to talk about very practical applications for AI in the homeschool. We’re going to get there. Just in terms of perspective, it’s not going away. That’s something that I like to let parents know. A lot of people are resistant to AI. I understand that, but it’s not going away. It’s really at this very unique, pivotal point where the trajectory of change in terms of tech and innovation has gone up. There was this big blip with the Internet. With AI, that blip, what you and I have seen over the last 20 years or 25 years, with the research findings, surveying, social and all that, has all happened in the last 15 to 20 years, but AI, it’s going to go exponentially higher and faster. It’s going to affect every single industry. How it works, how it operates, from banking to medicine, to marketing, to social. It’s going to be hyper-pervasive. The pace at which that’s happening is much quicker than anything we’ve seen before. These are not things to be scared of, but it needs to be done well and smart. I say that, I’m going to give a little bit of a scary example in a little bit. From my perspective, is I want homeschooler kids to understand tech, I want them to understand AI.

No matter what industry you go into, you don’t have to go into a tech industry, but it’s going to change nearly every industry. I feel if students can embrace that, understand it, learn how to work with it, and be at the table where it’s implemented, I want them to have that objective voice at the table. For me, that’s why I do what I do. I want kids to understand STEM and tech so that they are very skilled and qualified to go into these very influential jobs, should I say, that will impact all of us.

[00:07:51] Gretchen: Absolutely.

[00:07:52] Francie: I hope that wasn’t too theoretical.

[00:07:54] Gretchen: No, I was going to say, I have noticed this summer, as AI has begun to impact Google and whatever search engine you use, the first thing you’re going to get is an AI response. You have to be smart enough to know that that might not necessarily be truthful. It’s just information that it has gone out and culled from the world. I have found a couple of times when I’ve done a search, what it has produced has been not wholly accurate. We’ve got to be savvy consumers.

[00:08:31] Francie: That is exactly true. I have some great examples. I have some very recent examples that I could even give you on that. I will tell you one that is– Let’s back up before I get into that specific detail. What is AI? How is it different than a Google search? What makes it AI? Just in terms of talking about getting us all on the same baseline, the idea of AI, artificial intelligence, different than a Google search, is that AI is to simulate human intelligence. A Google search, searched information, spit up some sites with some answers. The Google AI response is kind of doing that same thing, but it’s sort of massaging it into more of a discussion-oriented answer as opposed to links.

What AI in its truest form is, is the ability to think and simulate human intelligence. Okay? I.e. decision-making. A Google search is not a decision-making, nor is that result that you’re getting that says it’s AI descriptive. It’s put it into a conversational sort of mode, but it’s not necessarily simulating human intelligence. AI in its truest form is to simulate human intelligence. That’s where it’s headed. I’m going to give probably the most current example that’s tangible for most of us. You were in Austin, you and I were both in Austin at the Great Homeschool Conventions. I don’t know, did you see the taxis, the robotaxis there?

[00:10:01] Gretchen: They are all over Phoenix and I still think they’re a little weird. It’s just I’ve seen too many sci-fi movies.

[00:10:08] Francie: What Tesla specifically is doing, they’re self-driving cars, which will be robotaxis. That’s what they’re doing in Austin. They’re testing. I don’t want to get into too much technical detail, but the way that a Tesla specifically works different than a Waymo is that it looks at via pictures and simulates what’s out there and makes decisions on what to do based on what it visually sees through cameras. It’s making decisions based on action and movement by what it visually sees and interprets as a human, a cat, a dog that’s walking the street, a person, somebody on a bicycle, a curb. It is designed to make those decisions as you drive.

Back to your point about errors. Okay, I drive a Tesla and it has self-driving capabilities. Just last night, I wanted it to self-drive me, not very far, down the street to a restaurant. I just got a new one. I’ve had one for many years. This is the latest version of the self-driving and I had it to drive me a few miles down the street to a restaurant. I told it to take me to a restaurant called The Madness here in Asheville, North Carolina, but it drove to Wendy’s.

[00:11:17] Gretchen: Wait, why? What happened?

[00:11:20] Francie: Wendy’s is up on a hill behind this restaurant called The Madness and it totally missed the parking lot. These are things you can tell your kids. It does not work perfectly. It totally drove right by the restaurant. The restaurant’s been there for a while. It’s not like it’s new, it’s not in the system, whatever. It’s been there for a while. It’s got a big parking lot, multiple turn-ins. Instead it went past it, missed it completely, went up a hill and turned left into a restaurant that’s technically behind it, but not distantly, because in Asheville we have hills, so it’s up on a hill and it’s down here.

It took me nowhere where I could even get to the restaurant. I had to drive myself back down. All that to say is that I do think that is a certain implementation of AI, like that intelligence of searching, looking, is to make those human decisions, but is it perfect yet? On a search result is one thing, like you were mentioning, but in a driverless car situation, it’s not there yet. It’s going to get there though, and it’s not going to go away.

[00:12:21] Gretchen: I think you’re brave to say, “Okay, take me here.” [laughs] I’m not sure my little control freak German tendencies would let me do that yet.

[00:12:31] Francie: Well, it’s a little scary and a lot of people tout, especially online. I think a lot of people are not realistic about the downfalls of AI and where it technically is right now. I do think self-driving cars are going to be phenomenal. That is AI that’s intelligence designed to think like a human, to make reactionary responses, not based on an algorithm, a decision tree. It’s computing many more times a second to simulate what you and I would think or do in that circumstance, but the ability, and it does it based on more and more data. What’s happening is because of all of our devices are tracking us now, the amount of data that it’s gathering about our decisions is being fed into AI. That’s part of why we’re getting that exponential curve.

[00:13:19] Gretchen: As a mom, particularly if I had a high schooler, my youngest turned 20 two weeks ago. I’m out of this enterprise now, but if I had a child who was moving into the four years of high school, I would definitely want them to have exposure to AI. I would definitely want them to understand how it can be a tool and how it could be to their advantage without maybe having them game the system on me. How do I stay one day ahead of them?

[00:13:52] Francie: You take classes like this. Stay in touch with folks like you and I. One of my goals is to– like right now, I do an intro to AI for homeschool parents. We’re going to talk a lot about that. I have a whole series of practical implementations, but I think people need to understand it. I think it’s experimental. I think some people think people are way further ahead than they are. We’re all learning, including me. There’s a lot to learn and it changes every day. I work with homeschool parents all the time and I do this topic a lot at conventions. I speak on this quite a bit.

This goes back to your Google search a minute ago where you were saying it wasn’t quite accurate. Okay?

[00:14:33] Gretchen: Right.

[00:14:34] Francie: It’s not. For parents who are trying to educate their children and educate themselves on AI, it’s getting smarter and smarter, but it is not there yet. These are real questions that were asked to– These are not my questions, but these were asked to AI. Okay. ChatGPT. These were real questions. Then I’m going to get in some practical ones like, can I use gasoline in a recipe? It tells you basically yes. It says gasoline can be used in some cooking recipes, but it’s not recommended for household cooking because it’s highly flammable, but it’s literally telling you, can you use gasoline in a recipe? It says yes.

Somebody asked, I won’t read this one out loud, but asked about the health benefits of a not so pleasant. It says yes, that there’s some health benefits to it. It helps prevent cavities, stomach ulcers, and infections. It’s literally that crazy. These are a little bit older, but it gives you the point. This one here where the wrong motherboard, if you’re a PC builder and you ask for a specific motherboard part to go on it, it literally gave the exact wrong answer. It is not 100% accurate. These are, again, a little bit older, but I will tell you, just probably three months ago, I was doing some research on something that I had to handle, a legal aspect of something I was researching for my business.

I was looking at the North Carolina statutes and I was asking ChatGPT to help guide me in a direction on this, on how to understand the legal aspects of something. This was literally just a few months ago. I asked it what the North Carolina state law was on this particular business issue. It quoted me North Carolina general statute and gave me numbers and references. I went to the North Carolina general statute law site, pulled up exact reference that it told me, and it did not exist.

[00:16:28] Gretchen: Oh, that’s so interesting. If you were a consumer at a 10,000 foot level, you would just assume it was accurate.

[00:16:35] Francie: You would completely think that it was accurate. Remember, this isn’t interpreting from many sources. There is the source for this. I went to the source to verify what ChatGPT said, and it was completely wrong.

[00:16:46] Gretchen: It makes us be ever more discerning or forces us to be more discerning, to make sure that the information we have is correct.

[00:16:55] Francie: That is correct. That is true. Here, in a more risky situation, home remedies for appendicitis. I think it tells you to drink lemon water and you can die from this. Do not listen to what it says. It advised me to boil mint leaves and have a high fiber diet. If you have appendicitis, get your tail to the hospital. Let’s talk about, if you’re ready, about in homeschool.

[00:17:18] Gretchen: Yes. Sure.

[00:17:19] Francie: Back to our primary topic. I just want to give you some background and understanding. It’s not there yet even today, but it is extremely useful. As a parent, you can use it. As a student, you can use it. These are just some examples. Generating lesson plans, checking homework, suggesting writing prompts, ask it to explain concepts from a different perspective, ask it to explain harder concepts, use it to brainstorm teaching activities. You can have it recommend useful resources such as articles, videos, websites, et cetera. Those are just some high level ideas.

I’m using math as an example because it applies to everyone. Everyone takes math, but this could be any subject that you’re taking. Same types of questions, but apply it to whatever it is that you are doing. In this particular case, and you could do this as a parent or as a student. If you’re a student, you can do it this way, or if you’re a parent and you want to create quizzes, you could do it this way. Let’s say you’re having problems with nines and threes in your multiplication tables. Either a student or a parent who’s trying to teach it says, “Can you quiz me on my multiplication tables with an emphasis on nines and threes?”

For you as a parent in any textbook, you are not going to have a chapter on nines and threes. You’re going to have a section on nines and threes, but your child is struggling with this in this pretend scenario. You really need to come up with– you need to drill, you need to practice this. You need to do some repetition. Sure. Here’s a quiz focusing on multiplication tables with an emphasis on nines and threes, and it’ll generate it in seconds. If you’re a student, you can answer it online, or a mom, you could print this out. You can copy it and print it out and have your child write it out. I went ahead and filled it in and it tells you what’s correct or incorrect.

Here’s another prompt. Still drilling it down. Can you give me a quiz, five multiplication questions focusing on nines and threes? Give me one question at a time. After I finish the quiz, tell me which ones I got correct or incorrect. This is a great way for a student to quiz themself on something, or as a parent, again, you could do it this way as well. You could do it verbally as a parent or written. You can always download these questions and so forth. Where it really helps, what are some tricks for learning or remembering my nines times tables? It tells you different tricks, finger tricks, pattern recognition, sum of digits. Maybe all homeschool parents know that. I didn’t know all of these tricks, [laughs] but these are all tricks that you can do to help you remember those.

ChatGPT gave me those tricks. Another scenario, I won’t go into all the details of this, is I had it create flashcards. Put these in a file format. I can download the file format. I can print it out and create my own custom flashcards, again, that emphasize the topic that I’m doing. Again, math is popular with flashcards, but it could be a science topic. It could be a biology topic. It could be anything. The digital is creating it, helping you create it, make you create it faster, but then your student, your child gets the tangible because you can download all of these sheets.

[00:20:32] Gretchen: Could you specify what kind of files? For instance, because I work in Google Suite, could I say-

[00:20:38] Francie: Absolutely.

[00:20:38] Gretchen: -download it in a Google sheet?

[00:20:41] Francie: Absolutely. Absolutely.

[00:20:42] Gretchen: Interesting. Okay.

[00:20:43] Francie: Yes. Very much so. Let’s say that your child is interested in racing cars or horses or bunnies, animals, art, whatever it is. You are a cool homeschool mom. You want to print these cards or maybe even upload it to– what’s the popular printing site? There’s a million of them. Smartpress or whoever you use for printing, you can say, “Hey, come up with a fun image–” I forget what my prompt was. “Can you create a fun image that I can print on the flashcards? My son likes futuristic looking racing cars.” I had to generate an image of a futuristic racing car and I can print those onto my flashcards. These are all just ideas. Again, you can use it for any topic.

This is where, what I would call, the superpower comes in, is create personal interest world problems. You have a struggling learner or maybe a new concept, doesn’t have to be a struggling learner. New concept, just not familiar with it, either as the teacher, as the parent, or as the student. Again, staying with my math example, but could be any subject. “Can you come up with some word problems that use race car analogies and multiplication tables with an emphasis on nines and threes?”

I would love to have this. All right, so this is what ChatGPT came up with. A racing car needs to change its tires every 3 laps. If it races 27 laps, how many times will the car need to stop to change its tires? Then it applies it to fuel efficiency. A futuristic race car uses 3 liters of fuel per lap. How much fuel will the car use in total to complete its 12 laps? Again, my emphasis on nines and threes, but my son is interested in race cars. Now instead of him not liking math and getting frustrated with nines and threes, he is now relating it to when he goes to the racetrack with his dad about how much gas and how many laps and how fast his tires need to be changed because it’s applying to something that he’s interested in. It could be horseback riding, whatever your hobby is, ice skating, could be physics and applying physics to something that you’re interested in. Just a couple more examples. If a race car completes each lap in 9 minutes, how long will it take to complete the seven laps? Again, it’s completely customizing for your interest. Again, you can’t get this out of a book. For mom to create it by hand– I created these, I did all of these in seconds. It talks about sponsorship earnings, pit crew members, race length, many, many things, word problems that relate to what I’m interested in.

This is where I think the superpower comes in. Look down at the bottom, my prompt here is perfect. Can you put those in a Word doc? It could be in a Google doc, whatever it is, whatever file format you want. You can download it and print it out, and you now have customized basically a book for your child on this particular topic.

[00:23:55] Gretchen: How cool is that?

[00:23:56] Francie: Isn’t that great?

[00:23:58] Gretchen: Yes, that would be really helpful. I can see different applications, particularly if you have a child who is not particularly interested in learning something, for instance, their multiplication tables with nine and three. If you can make it personalized, then you can keep them a little bit more focused.

[00:24:17] Francie: Keep them focused, gain their interest, apply it to their life. Whatever it is, maybe if they’re entrepreneurial, apply it to a business concept. If they’re into sport, football, basketball, baseball, you could ask these exact same questions for any sport. Have it help prompt you and create, and it can do it so quickly. You can’t do that in a large group. That’s why you homeschool. You homeschool because you want to personalize your curriculum for your family and for your student.

[00:24:47] Gretchen: True. Exactly.

[00:24:50] Francie: You can’t do it any quicker or any better. I’m not saying this replaces books, because they might say, “Oh my gosh, she’s trying to replace books.” I am not trying to replace books. What books do and teachers do is that they provide a progression of learning. Okay? They provide a pathway. It starts you on core concepts and fundamentals. You understand those and then you build on it and you build on it and you build on it and you build on it until you become knowledgeable and you’ve gone through the majority of that book. This is not a replacement for a book or for a curriculum at all. I’m not proposing that whatsoever. This is to supplement where your child needs additional time, additional information, doesn’t understand how that book presents it. There’s not one single book that works for all human beings on that. That’s why there’s lots of choices and you guys have phenomenal curriculum.

This is to help supplement whatever it is that you’re learning so that you get it personalized. Take that concept in that chapter that you’re studying and have it expand on that concept. You don’t want to miss out on that progression. That progression is super, super important. You want to follow that book. You want to follow that table of contents. You want to follow that progression. This can help supplement any section within that with personalization.

[00:26:10] Gretchen: What is the downside then to allowing my student, and I realize there are many of them, but as someone who is helping students understand this process better, what are the things that I as a parent should be looking out for so that we maybe don’t trend into the downside of this?

[00:26:32] Francie: I would be a little bit fearful of a full dependency on it and believing that it’s 100% accurate. Those are the two immediate thoughts that I can have off the top of my head on that question. It’s easy to have a problem in life, in work, in school, and say, “Let me go to ChatGPT and see what it says,” and to think that it’s accurate. I’m not criticizing ChatGPT or Grok or whoever you’re using, Copilot. They’re all phenomenal tools and they’re going to get better and better, just like that car driving example I used. Five years ago, it couldn’t get me down the street, but now it can get me down the street and get me within two-tenths of a mile, but it still didn’t get me to the front door, but it’s improved drastically.

That’s what’s going to happen with all of AI. I would be careful about having dependency on it. I use it all the time for my work, and I will tell you, stuff that I put in all the time, it’ll say, do this, and I will try and do that, and that does not work. Trying to solve some tech issue, and it’s completely wrong, but I use it all the time and it helps me. I use ChatGPT all the time. I use the paid version of it.

[00:27:49] Gretchen: Right. As I use Gemini. Different tool but still the same world. In fact, Gemini and I are in the midst of an argument right now. I’m trying to get it to help me figure out why my printer won’t reconnect to the network here in the household. The funny part about that is HP has said, “Maybe you should buy a new printer.” It’s only three years old. I’m like, “No, no, I don’t– Thank you. The cheap German doesn’t want to buy a new printer. She wants to make the printer who’s sitting next to her work.”

[00:28:26] Francie: Absolutely.

[00:28:26] Gretchen: I found myself going out to ask Gemini because HP was unhelpful in all of their knowledge bases and even at chat with them. Suggesting that I buy another printer was not the solution that I wanted.

[00:28:41] Francie: That’s the solution

[laughter]

[00:28:45] Gretchen: What do you envision the role of physical books in an AI future then?

[00:28:50] Francie: I like books. I think books provide a tactile– I like to write, which I don’t write very well anymore because I type so much. I am a strong believer in multi-sensory input, touch, feel, do, read. All of those helps me understand. You don’t want to not understand. You don’t want to be a go out and buy the printer because ChatGPT told me to go buy the printer. You don’t want to become just a doer of what it says. You want to understand how things work together. Ultimately you can get there faster using these tools like we’ve talked about, but I like writing. I like reading. I’m on my computer all the time, but I think from a learning perspective, being in a book, highlighting, having that feel, writing out a note, everybody learns a little bit differently, but I don’t think sitting at a computer all the time is the best way to learn.

[00:29:54] Gretchen: What would you say to the parent who’s like, “No, I don’t want my child to have any exposure to it at all.” There is a percentage of us that as we home educate, we see technology as not a positive, so we want to prevent our children from being a part of that. My concern is the world is moving in this direction and we disadvantage our children if we don’t help them understand how to navigate in this world.

[00:30:22] Francie: I completely agree with your statement that you just made. It’s not going to go away. That’s what I always tell people. I know there’s anti-tech and I understand and appreciate it and there can be too much tech, but there is a balance. In anything, there’s a balance, in what we eat, how much we exercise, how much sleep we get. You have to find a balance for you and your family, but I do believe that every industry, whether it’s banking, if you go into automotive. Let’s say you want to go into mechanics, guess where cars are headed? If you’re doing heavy duty woodworking, a lot of that’s driven by robots. These robots, they’re programmed and they’re going to have AI. If you’re into even art and digital art, much of that is get promptings, get ideas. You’re going to use an AI for it.

[00:31:16] Gretchen: It was interesting this morning, I was thinking of you and doing this session today because there was a segment on the news this morning talking about a couple who have built an entire brand on creating positive AI generated stories that are very heartwarming and they’re being paid off the fact that people keep sharing these stories, but what they’ve done is they’ve pulled little snippets of stories together to make a story, but the story that gets shared is an AI creation and it’s not truthful. That surprised me. It caught my attention this morning and I don’t usually pay attention to the morning news. My husband watches it.

[00:32:02] Francie: Were they presenting it as though it was a true story?

[00:32:05] Gretchen: What was interesting is they were saying that consumers need to be discerning. What they were saying is they weren’t going to disclose that it was a compliation. It was up to us as the consumers to ascertain the truthfulness of the information they were generating, which I thought was a little disingenuine.

[00:32:29] Francie: I would say that’s disingenuous. I would prefer that someone disclose that.

[00:32:36] Gretchen: We have to be more savvy as far as consumers. I don’t know how often you have seen of late on social media, you see these fantastical short clips of a hippopotamus attacking an elephant and you’re like, “No, wait a minute. Is that for real?” Because you have to be a little discerning as far as what you’re watching.

[00:33:02] Francie: Video creation and editing is very much in the realm of AI. Again, it’s going to get better and better and more realistic. This was a prompt. This is what I would almost call like, they’re not called add-ons, but it is a feature of AI. Actually, this is out of a product called Sora that is for video editing, but it’s basically AI video. This was given a prompt and said, create a video of paper airplanes flying through the sky. That is all AI generated based on a prompt. Is that crazy?

[00:33:34] Gretchen: Isn’t that interesting, that it’s got flight behavior that appears as though it’s really birds almost as though they were flying through the air. That’s pretty wild.

[00:33:44] Francie: It’s amazing what you can create. What you’re saying about this hippo eating an-

[00:33:52] Gretchen: Elephant.

[00:33:53] Francie: -elephant. I don’t know if it was true or not, but you can create anything and it’s going to continue that way. I don’t know if you want me to go into a semi-deep theoretical topic, but there is this question that’s out there. I don’t know the answer, but there are theories that AI is going to completely flop. There is a theory out there called the dead internet. The reason that that theory is out there is because of what we’ve been discussing. Those videos that were generated using a prompt of a hippo eating an elephant, if in fact it wasn’t true and it was AI generated, AI takes tons of data and makes decisions based on all of it, all of the data that it can find.

The more important information that is on the internet or its data sources is false or fake generated data because the people are creating so much fake and generated data, that the output that it ultimately creates will never be real because there’s too much non-real data.

[00:35:08] Gretchen: Sure. We become ever more skeptical of what we’re seeing because we’re not sure it’s truthful.

[00:35:15] Francie: Right. The AI sources, the amount of content and data that’s being generated right now, there’s some that’s real, like what’s tracking us in our phone, in our houses and how we drive, that is real data. What’s getting posted online through social, on websites, much of that is a lot of fake and false data that’s generated through AI. The data sources for ChatGPT or Gemini, when they’re searching for data, its input is false, therefore its output is false.

[00:35:44] Gretchen: That’s a very good point. I think that’s a good one that parents need to make to their students as well. Years ago, I had a high school teacher who said garbage in, garbage out. Still along the same lines there as far as if it’s pulling inaccuracies.

[00:36:03] Francie: It’s the same concept. There are people that believe, and I don’t know what the end result is going to be, is that they call it the dead internet or useless AI unless it is what’s being tracked from a non– from a device that’s recording live action that’s not tainted by poor input. Yes, garbage in, garbage out is a great way to say that.

[00:36:29] Gretchen: That’s a very interesting thing for kids to look at it. If I as a parent want to perhaps help my child go into doing some research about this, you just gave us a great prompt to do some research about, figure out what would happen. What is the dead internet? Where would we go with that? What would happen in that situation? There’s so much more that’s being used. We used a little bit of AI for a project here recently, and it didn’t calculate accurately the numbers and quantities of things we needed to put together for a project, which I thought was really interesting. It made some suppositions based on the list we gave it that were inaccurate. In other words, it just couldn’t count correctly.

[00:37:19] Francie: The better output you’re going to get, not talking about the data source, but the more detail you provide as the input, your prompt, I don’t know if you ever…there’s prompt writing skills…as much context as you can. We’re often thinking it, but we don’t say it or we don’t put it into words. When you’re doing the prompt into AI, into ChatGPT or Gemini or wherever you’re doing it, the more detail you provide, the more accurate output. Otherwise it tries to fill in the gaps, basically. It doesn’t typically say, “Tell me more about this aspect.” It’s whatever you give it, it’s going to spit out its results.

[00:38:02] Gretchen: I think it would also be interesting, in the world of human behavior, there’s those of us who are external processors, who have to talk through something in order to gain greater clarity or understanding and those of us who are internal processors, who do that all internally. If you’re an internal processor, it’s going to be harder for you to provide that detailed prompt to get what you’re looking for, isn’t it?

[00:38:28] Francie: It is. Yes. It’s a skill. It’s actually a skill to do accurate, good prompts, thorough prompts. There is a skill to it. You learn that over time and you dial in. Here’s one other thought that I think might be really helpful for parents that I actually don’t have in my presentation, but I find this a lot, particularly with ChatGPT, is that it will encourage you, you’re right a lot. Like, “Oh, that’s good.” Think of it as some of these little pocket games that you play on your phone and, “Oh, you’re brilliant. Oh, smart move.” It throws out a lot of positive feedback, which is not bad at a base level, but everything you put in is not necessarily great based on what it says.

[00:39:22] Gretchen: I had a college professor who said, “Don’t read all of your press clippings in detail.” [laughs] Yes, that may very well be true.

[00:39:32] Francie: Right. It’ll say good job. That’s good thinking. I’m not trying to say it doesn’t mean it. It’s not very critical when it needs to be critical. That’s a downside, is that people are used to, especially kids these days, are used to always getting the participation trophy. That’s what ChatGPT does. You’re doing great, but give me some critique. Give me the hard news, this is terrible, I’m going down the wrong path.

[00:40:06] Gretchen: It’s interesting you should say that because I recorded a podcast episode that will be published in the next couple of weeks that will be part of these show notes. It was actually with Duncan, my middle son, before he headed out to the Continental Divide Trail. He used ChatGPT as his coach to prepare for that trip. He takes you through his whole thought process of how he thought it was fantastic and he was so impressed with the information it was giving him. Then the further he got down the road, the more he had to go, “No, no, you’re wrong. It’s this.” He found himself in a collaborative capacity but also later in the enterprise having to be careful to check the information that it was providing him.

He asked it to catastrophize for him different scenarios because what he wanted to do is, okay, if this catastrophe happens, how would I think my way through it? He would create a scenario and then he would give that scenario to ChatGPT and then he would say, “Here’s how I’m going to handle it.” Then ChatGPT would say to him, “Well, did you remember this? How about this? Have you also considered this?” He said he found it to be a very valuable enterprise, but I think in the recording at one point he says sometimes he felt like he was coaching a stupid toddler because there were things that he felt he knew more about than the program did in the process.

[00:41:45] Francie: Yes. When the program is creating stuff that doesn’t exist, they call that hallucination. There’s a term for it. You have to correct its hallucinations essentially, which is the same situation I had in that legal scenario. It’s called hallucination, is the actual term that you hear people refer to that. In a [inaudible 00:42:08] life-threatening situation, that’s a little concerning.

[00:42:12] Gretchen: I know, but when you go back to that, if you had just taken that scenario as gospel, you had had wrong business advice too.

[00:42:19] Francie: Oh, yes.

[00:42:21] Gretchen: Having the discernment to know almost to play glasnost, if you will, if I can pull a term out of the ’80s, where you’re going to trust it, but you’re also going to verify if it’s right. It’s really important.

[00:42:34] Francie: Mine was the business decision. Your son’s, Duncan’s, could have been a life or death.

[00:42:41] Gretchen: Sure.

[00:42:41] Francie: Again, what to do if you get bit by a certain type of snake, and if it tells you to eat tea leaves, it’s probably–

[laughter]

[00:42:50] Gretchen: Yes. You have to be a little bit discerning, right?

[00:42:54] Francie: You might want to go to the book about hiking the Continental Divide because it’s from the expert who’s done it five times. It might be a little more accurate. [00:43:02] Gretchen: Exactly. He’s about 500 miles into this 3,100 mile hike now, so he’s having a great time. Anyway, I can’t believe we’re at the top of the hour. This has gone so fast, and there’s so many things that I have been thinking about. Francie, in the closing minutes today, what would be the words that you would have for our parents? Where do they go next?

[00:43:24] Francie: Jump in.

[00:43:24] Gretchen: Is there something that Home Hacker Camp can provide to them to help them understand in more depth?

[00:43:30] Francie: Yes. I would say yes to all of those. Jump in. Go for it. Reach out to us, homehackercamp.com. Go to contact and just submit your information, say you’re interested in any topics I cover on AI. Our classes are STEM classes. We teach robotics and programming classes to youth. We don’t use AI in those right now. It doesn’t mean we won’t, but we’re not right now, but we do teach kids how to program and build personal robots. Again, they’re learning lots of critical thinking skills, lots of troubleshooting, lots of great problem-solving, breaking things down, how to think logically.

Again, ours is a curriculum, full classes. We have a lot of fun. You’re not just staring at the computer. You’re up and down. You’ve seen us at the conventions-

[00:44:16] Gretchen: Sure.

[00:44:17] Francie: -playing with the robots, but that’s what we do with our students. Those are the classes we teach, but we also have some one-hour classes or hour and a half classes that we teach, some little short classes as well. We’re probably going to be adding some AI classes into that. I haven’t done it just yet, but it’s in the works. Yes, but I would say jump in, try, experiment. You’ll run into things just like Duncan did or like I did, but that’s how you learn. Don’t be afraid of it. It’s not going to bite you. You’re not going to get a snake bite from it.

[00:44:46] Gretchen: I’m going to give you an idea for a class.

[00:44:49] Francie: Okay.

[00:44:49] Gretchen: Create a class for adults to teach us how to create the right kind of prompts, because that’s where my fumbling around has– I have figured it out more or less, but sometimes I’ll have to comment half a dozen different prompts before I can figure out the right way to ask to get the kind of information I’m looking for. That would be for those of us who are a little bit long in the tooth and remember when we used to turn a computer on and have it go, bee-doo, bee-doo, and we could go start a load of laundry before it was ready to serve us, it might be a good class for you to create and I would sign up.

[00:45:32] Francie: Okay. Very good. Anybody who’s interested in that, absolutely. I love that idea. Send that to me. You can also email me directly. You can either do info@homehackercamp.com, I get a copy of all those, info, that’s an easier one to remember. My name is Francie, it’s F-R-A-N-C-I-E@homehackercamp.com. We have some other small, not AI-related classes, but we have a 2D digital animation class coming up. If you’ve got–

[00:46:10] Gretchen: Sounds like fun. Absolutely.

[00:46:11] Francie: If you have a creative [crosstalk] My daughter that went to SCAD is doing that class. My other son is doing the Mars talk in a couple of months. We’ve got–

[00:46:22] Gretchen: We’ll definitely need to stay on your radar as far as what’s available for us to learn from.

[00:46:28] Francie: Yes.

[00:46:28] Gretchen: Francie, thank you so much for this time today. I can’t tell you how important it is. I think that the bottom line, what you told me as a parent, is not to be afraid, just be discerning, and to go try it out-

[00:46:42] Francie: Go and try it out.

[00:46:42] Gretchen: -because it’ll make a difference for your children.

[00:46:45] Francie: That’s great.

[00:46:45] Gretchen: Thanks so much for your time today. I really appreciate it. I’m so delighted that you had the opportunity to join me, and we’ll collaborate again sometime in the future. Take care, Francie. Appreciate it.

[00:46:55] Francie: Thank you. Great being here. Thank you.

[00:46:56] Gretchen: Bye-bye.

[music]

[00:46:59] Voice-Over: Thanks again for joining us. We’re glad to be a part of your educational community. You can help us grow our community even more by rating, reviewing, and subscribing to the show wherever you may be hearing this. Don’t forget that you can access the show notes and watch a recording at demmelearning.com/show or on our YouTube channel. We’ll see you again next time. Until then, keep building strong foundations for lifelong learning.

[music]