Latest / Dr Justin Coulson's Happy Families / AI Is Making Our Kids Dumber and Schools Are Letting It Happen

Transcript

AI Is Making Our Kids Dumber and Schools Are Letting It Happen · Dr Justin Coulson's Happy Families

0:05
Speaker 1: Hello and welcome to the Happy Families Podcast. Two months ago,
0:09
Speaker 1: the headlines screamed, chat GPT is destroying our brains. AI
0:15
Speaker 1: is lowering your IQ, We're all getting dumber. In fact,
0:19
Speaker 1: one Australian newspaper article said that education is no longer education.
0:26
Speaker 1: We have been lobotomized by AI. The alarm came from
0:31
Speaker 1: an MIT study that showed that students who used chat
0:34
Speaker 1: GPT had terrible recall. Eighty three percent could not remember
0:39
Speaker 1: what they'd written, compared to just eleven percent who couldn't
0:42
Speaker 1: remember if they used their own brains. Case closed rights.
0:45
Speaker 1: AI makes you stupid, Except there's something buried in that
0:49
Speaker 1: two hundred page study that most headlines missed, and it's
0:52
Speaker 1: something that changes everything today. What AI is actually doing
0:55
Speaker 1: to our kids' brains, why schools are accidentally making it worse,
1:00
Speaker 1: and the one thing that you need to teach your
1:02
Speaker 1: children before they touched chat gpt again. Stay with us, gooda.
1:11
Speaker 1: Welcome to the Happy Family Podcast, where you get real
1:13
Speaker 1: parenting solutions every single Day's Australia's most downloaded parenty podcast.
1:16
Speaker 1: We are Justin and Kylie Coursen, the parents of six
1:19
Speaker 1: children who do from time to time use chat, GPT
1:21
Speaker 1: and other forms of AI. And I'm not just talking
1:24
Speaker 1: about the kids.
1:25
Speaker 2: I was recently having a conversation with my nutritionalists and
1:28
Speaker 2: asking him whether AI was a good way for me
1:30
Speaker 2: to count calories right, and he said it's not very reliable.
1:35
Speaker 2: And as we continue discussing, I never thought of this,
1:39
Speaker 2: but people will take a photo of the plate of
1:41
Speaker 2: food they're eating and then ask AI to calculate the calories.
1:46
Speaker 1: That can't be accurate.
1:48
Speaker 2: It's not no, it's not, he said. He said, it
1:50
Speaker 2: will do. It will do a fairly good job, but
1:53
Speaker 2: like if you're if you're really serious about it, it's
1:56
Speaker 2: not going to come close. Yeah.
1:57
Speaker 1: AI can do stuff, but it can't do everything.
2:01
Speaker 2: Well, I never thought of that, and so I said, no, no, no,
2:03
Speaker 2: that's not how I'm using it. I said, I'm actually
2:05
Speaker 2: weighing all of my food and then I'm asking it
2:07
Speaker 2: to give me the information I'm looking for. And he said, oh,
2:11
Speaker 2: you're actually you're actually doing a great job. Work first,
2:16
Speaker 2: but you need to do so. He gave me a
2:18
Speaker 2: good pat on the back. It does have its uses.
2:21
Speaker 1: You've got to do the work for it to work,
2:23
Speaker 1: which is what today's podcast.
2:24
Speaker 2: Is all about, So what is this study that you're citing.
2:26
Speaker 1: Okay, so there's some MIT researchers. They found that it's
2:30
Speaker 1: not whether you use AI that matters, it's when, so
2:32
Speaker 1: that ties in exact with what you're hinting at. They
2:35
Speaker 1: tested these two groups. One group used their brain first
2:38
Speaker 1: and then used AI, so they did the thinking, they
2:40
Speaker 1: did the planning, they did the work, and then they
2:43
Speaker 1: plugged it into chat GPT to refine the work. Whereas
2:47
Speaker 1: the second group they went, I'm going to use AI
2:50
Speaker 1: and then I'm going to try to think about what
2:51
Speaker 1: the AI is giving me. So chat GPT then brain
2:55
Speaker 1: the brain to AI group. Their EEG scans. Their brain
2:59
Speaker 1: scan showed increased neural activity, so there was better attention,
3:03
Speaker 1: there was greater planning, more memory processing. And then even
3:06
Speaker 1: when they use AI, their brain stayed engaged because they
3:09
Speaker 1: knew the work, they knew what they were doing. They
3:11
Speaker 1: used their brain first. But the AI first group that
3:14
Speaker 1: then used their brain they just never recovered. Even when
3:17
Speaker 1: they stopped using AI, their cognitive engagement flatlined. It actually dropped,
3:23
Speaker 1: like it just stopped working.
3:24
Speaker 2: And that correlated with the conversation I was having with
3:27
Speaker 2: my nutrition lists exactly. I said to my I don't
3:29
Speaker 2: want to use an app that's just about counting calories
3:32
Speaker 2: because I'm still learning. I'm still trying to understand what
3:35
Speaker 2: fuels my body and where I'm going to best to
3:38
Speaker 2: get those foods from. So when I eat something and
3:41
Speaker 2: it says to me, you know what, that's pretty dense
3:43
Speaker 2: in calories, right, I can then say to it, what
3:46
Speaker 2: would I do differently if I want to still feel satisfied?
3:51
Speaker 2: And it's been such a really good tool for me
3:53
Speaker 2: to use as I start to understand this process.
3:56
Speaker 1: Yeah, So how you enter the AI interaction will determine
4:02
Speaker 1: whether or not it improves and amplifies your thinking or
4:05
Speaker 1: whether it replaces and subsequently diminishes your thinking. There's a
4:11
Speaker 1: difference between thinking with AI and thinking by AI. And
4:18
Speaker 1: so Alfie Cone is somebody that I've had on the
4:20
Speaker 1: podcast a number of times. You've become a fan because
4:22
Speaker 1: you've heard him speak so well. He puts it like this.
4:25
Speaker 1: He says, using AI to write for you is like
4:27
Speaker 1: bringing a forklift to the gym and having it lift
4:29
Speaker 1: weights for you. The weights get lifted, but you don't
4:32
Speaker 1: get stronger.
4:34
Speaker 2: He also said, by rushing to make use of chat GPT.
4:39
Speaker 2: Schools may not just be overestimating the capacities and capabilities
4:44
Speaker 2: of artificial intelligent, but underestimating the essence of education.
4:50
Speaker 1: Yeah, I like African. The terrifying part for me is
4:52
Speaker 1: that schools are legitimately rushing into it. I read a
4:55
Speaker 1: recent investigative report in the Australia a couple of weeks ago,
4:58
Speaker 1: and oh, my god, goodness, this thing. University has changed
5:03
Speaker 1: a lot since I was there, is all I can say.
5:05
Speaker 1: They found up to ninety five percent of university students
5:09
Speaker 1: are now using AI to complete assignments and exams, and
5:12
Speaker 1: not just occasionally, I mean constantly. We've got a couple
5:14
Speaker 1: of kids who are at university they do their exams at home.
5:18
Speaker 1: Like there's software that monitors whether anyone comes into the
5:21
Speaker 1: room or with you achieving apparently, but they're literally doing
5:23
Speaker 1: their exams at home. One student, his name was Hayden
5:27
Speaker 1: in this Australian report, graduated with a high distinction and
5:31
Speaker 1: when he was asked how much of his final year
5:33
Speaker 1: he outsourced to AI, he said, and there's a direct
5:35
Speaker 1: quote all of it. One hundred percent of my coursework
5:39
Speaker 1: and one hundred percent of my exams, So he went
5:43
Speaker 1: from being a C average student to an A plus
5:46
Speaker 1: student in eighteen. No, he didn't, he didn't get smarter,
5:49
Speaker 1: he didn't become an A plus student. Chat GPT did
5:52
Speaker 1: the work for him. And I've got a couple of
5:55
Speaker 1: friends who are still University of the lecturers. I obviously
5:57
Speaker 1: am not lecturing at UNI anymore, but they basically said, yeah,
6:00
Speaker 1: we're not actually lecturers anymore. We are human detectors of plagiarism.
6:04
Speaker 1: But there's no point pointing out the plagiarism because everybody
6:07
Speaker 1: is doing it.
6:09
Speaker 2: Yeah. Yeah. A conversation I just recently had with someone
6:12
Speaker 2: was the acknowledgment that our schools don't pick up their game,
6:15
Speaker 2: that teachers literally lose their roles in the classroom because
6:18
Speaker 2: AI is doing everything.
6:20
Speaker 1: But here's the thing, though, these students are graduating and
6:22
Speaker 1: they are going on to become doctors and engineers and accountants,
6:25
Speaker 1: and they have degrees that they did not earn. You've
6:28
Speaker 1: got pilots who, well, I guess you've got to get
6:31
Speaker 1: into the plane and know how to fly the thing.
6:33
Speaker 1: But I think that it's terrifying. I just I think
6:36
Speaker 1: that it's really really worrying.
6:37
Speaker 2: So are we just talking about university study here.
6:40
Speaker 1: No. In that same report, in this train's analysis, they
6:44
Speaker 1: talked to a thirteen year old boy named Cameron. He said,
6:47
Speaker 1: in year eight, about twenty four out of twenty five
6:50
Speaker 1: kids in my class use AI for everything. My teachers
6:53
Speaker 1: don't blink an eye if I get one hundred percent,
6:56
Speaker 1: I don't even bother to fake wrong answers anymore. Oh
6:59
Speaker 1: my goodness. So I would say that schools are not
7:01
Speaker 1: just ignoring it. I would say schools are enabling it.
7:04
Speaker 1: You've got surveys showing that forty to sixty percent of
7:06
Speaker 1: teachers are now using AI to create their lessons, and
7:09
Speaker 1: get this, they're using AI to great assignments.
7:12
Speaker 2: So I was under the impression, based on what teachers
7:15
Speaker 2: have told our children in the past, that they actually
7:18
Speaker 2: had software that could determine whether or not their students
7:22
Speaker 2: had used AI and what percentage of their work was
7:26
Speaker 2: AI generated. Is that not the case?
7:30
Speaker 1: It is, but a lot of teachers aren't doing anything
7:32
Speaker 1: about it. They're just leaving it b and therefore nobody's thinking,
7:35
Speaker 1: noybody's learning. It gets worse as well. There was a
7:38
Speaker 1: study that was published in January last year January twenty
7:42
Speaker 1: twenty five in so the Lands, it's a good, high
7:46
Speaker 1: quality medical journal. They found that doctors who used AI
7:49
Speaker 1: tools regularly became less skilled within six months. So they're
7:53
Speaker 1: diagnosed as accuracy dropped when AI was taken away from them.
7:58
Speaker 1: They became over a line, they became less focused, they
8:01
Speaker 1: became less motivated to think critically. And so if that's
8:05
Speaker 1: what's happening to doctors, this is what's happening to our
8:08
Speaker 1: children's brains as well, turning to mush brain rot.
8:12
Speaker 2: Perhaps this is an obvious question, but why are kids
8:16
Speaker 2: doing it?
8:16
Speaker 1: If you're a student, why would you do it?
8:18
Speaker 2: Because I don't want to do the work.
8:20
Speaker 1: Yeah, So I think that's the number one thing I
8:21
Speaker 1: keep on talking to teachers in staff PD that I
8:23
Speaker 1: do around the country. If the work doesn't feel like
8:26
Speaker 1: it has any purpose, you can't expect the students to
8:28
Speaker 1: be motivated to do it, full stop, end of story.
8:29
Speaker 1: No matter no matter what you do, if students do
8:32
Speaker 1: not see some value, some purpose in the work, they
8:35
Speaker 1: are not going to have any desire to be there
8:37
Speaker 1: and to be doing it.
8:39
Speaker 2: So, for me, the hardest part of any project is
8:43
Speaker 2: actually starting getting started.
8:44
Speaker 1: Yeah, that's right. Literally yeap break through the Earth's gravitational
8:47
Speaker 1: pull like the Space Shuttle ten minutes of massive energy expenditure,
8:50
Speaker 1: and then you're in orbit and you're firing. But getting
8:52
Speaker 1: the kids out out of the atmosphere and into orbit
8:57
Speaker 1: hardest thing.
8:58
Speaker 2: And so a couple of times I've actually is AI
9:02
Speaker 2: to just generate a starting point for me, because I
9:05
Speaker 2: just I can't.
9:08
Speaker 1: That's that's an unusual thing, though, because usually once it
9:10
Speaker 1: starts for you, I'll give you another prompt like would
9:11
Speaker 1: you like me to do this next? And it makes
9:13
Speaker 1: you fairl I say.
9:14
Speaker 2: I say, well, I kind of like the gist, but
9:17
Speaker 2: this isn't me and so then I throw back my
9:19
Speaker 2: ideas and I'm like I like this better.
9:21
Speaker 1: Yeah, But the conversation continues and it feels like you're
9:24
Speaker 1: doing the work, but the AI is actually doing the work.
9:26
Speaker 1: You're just arguing with the AI. So I've used AI
9:29
Speaker 1: a lot over the last little while. I don't use
9:31
Speaker 1: it for my writing anymore. I just refuse to. I
9:34
Speaker 1: absolutely make sure that I've done the work first. If
9:36
Speaker 1: there's something really really important or something that I really
9:38
Speaker 1: need to refine, I'll ask it to do some refining
9:41
Speaker 1: for me. But I will not let it do the
9:44
Speaker 1: work for me anymore because it sounds like AI. And
9:47
Speaker 1: I've been really public about not wanting to sound like
9:49
Speaker 1: AI because too many people out there are using it
9:51
Speaker 1: and they you can't even read an opinion piece in
9:54
Speaker 1: the newspaper. Now it's all done by AI. Let me
9:56
Speaker 1: go back to why else they're using it? Everyone else
9:58
Speaker 1: is doing it. Not only is a purpose thing, but
10:01
Speaker 1: everyone else is doing it. And actually, I think schools
10:03
Speaker 1: are encouraging it. I think teachers are using it, administrators
10:06
Speaker 1: and pushing it. Tech companies are profiting from it. It's
10:08
Speaker 1: in big text interest for the kids to be using it.
10:11
Speaker 1: So after the break, a couple of quick minutes on
10:14
Speaker 1: what we can do the solution, what we're supposed to
10:16
Speaker 1: do to help our kids to navigate a new world
10:20
Speaker 1: of AI in a way that does not rock their brains.
10:23
Speaker 1: Stay with us. We're back. It's the Happy Families Podcast.
10:33
Speaker 1: If you're enjoying the pod, please like it, share it,
10:35
Speaker 1: leave a rating and review, tell people about it. It's
10:38
Speaker 1: how we get the word out and make families happier.
10:41
Speaker 2: So you're suggesting you have some solutions and some scripts both.
10:45
Speaker 1: I have both. The MIT study is really clear. Use
10:49
Speaker 1: your brain first, then use AI that works. Use AI first,
10:52
Speaker 1: then use your brain that doesn't work. So the rule
10:55
Speaker 1: that I think that everyone should have in their houses
10:56
Speaker 1: as follows. Before you touch chat GPT, you must attempt
11:01
Speaker 1: the problem yourself. We call this productive struggle. We call
11:04
Speaker 1: it developing distress tolerance or frustration tolerance. When kids struggle
11:10
Speaker 1: with things, when they write an awful first draft, when
11:13
Speaker 1: they try and when they fail, that's where the learning happens.
11:17
Speaker 1: And then, and only then can you use AI either
11:19
Speaker 1: to refine or challenge or extend your thinking, not to
11:22
Speaker 1: replace your thinking, to amplify it. Even then, I think
11:25
Speaker 1: that we want to be hesitant. I think we want
11:27
Speaker 1: to be really really careful with this. I'd rather the
11:30
Speaker 1: students do all of it.
11:32
Speaker 2: So what does this look like in practice?
11:35
Speaker 1: Let me give you a prompt. I've jotted down a
11:38
Speaker 1: prompt here. This is the recall prompt, so you could
11:41
Speaker 1: prompt AI as follows. Before you give me any suggestions
11:44
Speaker 1: on how I can improve this, I need to demonstrate
11:47
Speaker 1: my current understanding. So you tell the AI, you tell
11:50
Speaker 1: chat Claude and whatever it is Gemini, ask me to
11:55
Speaker 1: recall and explain my main thesis, the supporting evidence, and
11:59
Speaker 1: my biggest concern earns from memory, and don't give me
12:02
Speaker 1: any new input until I've successfully articulated these three things. Now,
12:06
Speaker 1: I don't think many students are going to do that.
12:08
Speaker 2: Well, that's a mouthful.
12:10
Speaker 1: You're basically saying, make me tell you what I'm going
12:14
Speaker 1: to do in detail before you give me any prompts,
12:18
Speaker 1: before you give me any new input. Before you let's
12:21
Speaker 1: make sure that I'm understanding.
12:22
Speaker 2: Then why would you bother?
12:24
Speaker 1: Well, you would bother because then you can do the
12:27
Speaker 1: second prompt that I've got, and that is the challenge prompt. Okay,
12:29
Speaker 1: so the first one's about recall. The second one is
12:31
Speaker 1: about challenge. I've written this essay. I've demonstrated that I
12:35
Speaker 1: know what I'm talking about here. I don't want you
12:37
Speaker 1: to rewrite it. I want you to ask me three
12:39
Speaker 1: hard questions that expose weaknesses in my argument make me
12:44
Speaker 1: defend my thinking. So if you use chat GPT or
12:48
Speaker 1: Claude or Gemini or any of the others like that,
12:52
Speaker 1: that's going to deepen and amplify thinking. That's going to
12:55
Speaker 1: make you better. It's like if I'm going to say
12:57
Speaker 1: I've got this exam coming up. Here's all the material
12:59
Speaker 1: that I need to hit me with ten questions.
13:01
Speaker 2: Yeah, but human nature is to find the path of
13:04
Speaker 2: least resistance.
13:05
Speaker 1: Not necessarily. Look at the way you're working out in
13:07
Speaker 1: the gym. You're not going for the path of least resistance.
13:09
Speaker 1: That's sitting on the cutch eating chocolate chip cookies and
13:11
Speaker 1: milk and watching TV.
13:13
Speaker 2: Yes, but I'm going against my nature because my nature
13:15
Speaker 2: would tell me to do exactly that.
13:18
Speaker 1: Nature or not, You're making a decision to be effortfull.
13:20
Speaker 1: And what I'm arguing is that the effort pays off.
13:22
Speaker 1: You're seeing the effort paying off and you're glad for it.
13:24
Speaker 1: Children students are going to struggle to see the effort
13:27
Speaker 1: paying off. They just see all their friends doing nothing
13:29
Speaker 1: and then hanging out on Fortnite or Roadblocks or Discord,
13:32
Speaker 1: and they're like, well, I just want to use Chat
13:34
Speaker 1: as well. I just want to use AI as well.
13:36
Speaker 1: This is going to damage their brains. So here's the
13:41
Speaker 1: script for the kids. I know skill feels like it's
13:44
Speaker 1: jumping through hoops. I know everyone's using AI. I know
13:47
Speaker 1: it's faster, I know it's easier, but you're trading away
13:51
Speaker 1: your ability to think. Right now, your brain is using
13:56
Speaker 1: developing neural pathways that you're going to use for the
13:58
Speaker 1: rest of your life. And every time you let CHAT
14:01
Speaker 1: GPT think for you, the pathways don't form. So then
14:05
Speaker 1: you can use Alfie Cone. It's like taking a forklift
14:07
Speaker 1: to the gym. The ways are getting lifted, but you're
14:10
Speaker 1: not getting any stronger. So just say to the kids
14:12
Speaker 1: you can use AI, yes, but brain first. Always no shortcuts,
14:16
Speaker 1: because what happens when you're a doctor and you can't
14:19
Speaker 1: diagnose without chat GPT, without the AI, it's going to
14:24
Speaker 1: be a disaster. That's the conversation to have with the kids.
14:29
Speaker 1: Low probability play. I don't think that it's going to
14:31
Speaker 1: work for everybody, but it's the best that I've got
14:33
Speaker 1: because this is a tough problem.
14:35
Speaker 2: What would you say to schools if you're struggling with
14:37
Speaker 2: this in the classroom?
14:39
Speaker 1: Oh, I would say to schools to go analog, not digital,
14:44
Speaker 1: and I would say teach exactly the same principle. I
14:47
Speaker 1: would also school leaders are unlikely to listen to parents
14:50
Speaker 1: around this. But what I would also say is that
14:52
Speaker 1: I would be demanding face to face exams don't let
14:54
Speaker 1: anything happen online with schools. That's likely to happen anyway, right,
14:57
Speaker 1: everyone's at school. It's not like they're doing exams from home.
15:00
Speaker 1: I would be demanding oral assessments, more kids standing up
15:03
Speaker 1: and learning how to articulate in front of an audience.
15:06
Speaker 1: I think it's one of the most important skills they
15:07
Speaker 1: can have anyway, Like for your student, your child, to
15:09
Speaker 1: be able to do that, that's a powerful skill. And
15:12
Speaker 1: we don't do enough oral stuff. And I'd be looking
15:16
Speaker 1: for assignments that do not just require text output like
15:20
Speaker 1: an essay or a lab report. I'd be looking for
15:22
Speaker 1: assignments that require a demonstration of human involvement. They would
15:28
Speaker 1: be the things that I'd be looking for. My sense
15:31
Speaker 1: is that a lot of teachers are just horrified by
15:33
Speaker 1: what's happening, but they feel completely and utterly powerless. But
15:37
Speaker 1: that's the standard, and I think schools should be held
15:38
Speaker 1: to it.
15:39
Speaker 2: So I guess, ultimately, just like all of the challenges
15:43
Speaker 2: that we have in family life, they're watching us, they're
15:46
Speaker 2: watching how we're interacting with things, and AI is no different.
15:50
Speaker 2: Whether it's social media technology in general or artificial intelligence.
15:55
Speaker 2: It's an opportunity for us to just take stock of
15:57
Speaker 2: how we're using it and ensuring that where modeling a
16:02
Speaker 2: positive experience for everybody.
16:04
Speaker 1: Yeah, here's what's at stake. We are raising a generation
16:07
Speaker 1: that may I don't want to sound too alarmist, I'm
16:10
Speaker 1: going to say, may never learn to think critically. That's
16:12
Speaker 1: probably going too far, but we're certainly handicapping their ability
16:15
Speaker 1: to think critically, solve problems, independently, engage deeply with ideas.
16:20
Speaker 2: But I don't think we can blame AI on that entirely.
16:22
Speaker 2: I think that that's been coming up for a long
16:24
Speaker 2: time in our school system based on the current model
16:27
Speaker 2: we're working with.
16:28
Speaker 1: Yeah, yeah, no doubt. The challenge really though, is let's
16:32
Speaker 1: get our kids thinking. So I would suggest this before
16:34
Speaker 1: your child uses AI for anything, whether it's their schoolwork
16:38
Speaker 1: or something they're writing, whatever it is, they just have
16:40
Speaker 1: to attempt to themselves first, brain first, then AI. That's
16:44
Speaker 1: the order always, and if they push back, just remind them.
16:47
Speaker 1: The forklift lifts the weights, but you don't get stronger.
16:50
Speaker 1: You choose do you want to think or do you
16:53
Speaker 1: want a machine to think for you? Because once you
16:55
Speaker 1: lose that ability to think, it's really really hard, to dive,
17:00
Speaker 1: to refine it, to build on it. Okay, that's it
17:03
Speaker 1: for today. Thank you so much for listening. We hope
17:05
Speaker 1: that this has been an informative and encouraging conversation. The
17:09
Speaker 1: Happy Family's podcast is produced by Justin Ruland from Bridge Media,
17:12
Speaker 1: with Muhammed's providing additional support. If you'd like more offer
17:15
Speaker 1: to make your family happy, visit happy families dot com
17:17
Speaker 1: dot au