Is AI making our brains lazier?

ABC. This podcast contains audio advertising. To hear it without ads, head to the ABC Listen app now. ABC Listen. Podcasts, radio, news, music, and more. It's one of Australia's most baffling, unsolved missing persons' cases. In 2007, four people, including a five-year-old girl, vanished in Western Australia, amid rumors of bizarre behaviour, cults, and long-hidden secrets. Yep, yep. We're trying to track them down. I'm Dominique Bans, and I've been investigating for the new season of Expans, the Nanup 4. Search Expans on ABC Listen or wherever you get your podcasts. The "Pace at Witch" developments in AI are happening is dizzying. Overwhelming. At the start of 2023, we were here. If you haven't, then get ready. GPT is a type of artificial intelligence model developed by OpenAI. So quite an innocent. Three years later, we are here. A stark warning about artificial intelligence from Jeffrey Hinton, the Nobel Prize-winning scientist. Hinton says there's a 10-20% chance AI will wipe out humans. Researchers have been sounding the alarm bell as they walk out to major companies. A slew of high-profile employees have quit OpenAI, and tropic and ex- That is quite the escalation. And along the way, a lot of other stuff has happened as well, from the peculiar to the deeply concerning. Although, maybe it's all deeply concerning, depending on where you stand. In Japan, one woman tied the knot with her AI-generated boyfriend. A number of reports about people developing distorted thoughts or delusional beliefs. There's a new social platform that's not for humans, but for AI bots to chat together instead. Two years ago, I had cognitive neuroscientists, Professor Joel Pearson, on the show, to talk about the coming AI revolution and the fact that our minds aren't ready for it. Joel is the director of the Future Minds Lab at the University of New South Wales, and he researches the psychology of artificial intelligence. And now, in 2026, the revolution is well and truly here. AI is no longer an abstract threat. It's causing job loss right now and the headlines they tell the story. So, I invited Joel back on the show for another conversation about artificial intelligence. And in this chat, we talk about his anxiety levels around AI, what it could be doing to our brains, hint, cognitive atrophy, and tips for keeping yourself human in an artificial age. I'm Sana Kadar from ABC Radio National. This is all in the mind. All right, so it's been nearly two years since we last spoke. First of all, how are you? I am well, all things considered. Thanks, Sana. Yeah, it's been a while. It's been an infinite time in AI times. I know. I know. And more to the point, really, what I want to know is how much more or less anxious are you now about the impact of AI on our psychology than two years ago? Probably about the same level of anxiety, slightly different in different ways. I've been working on trying to be more optimistic. Right? So, I think I'm a tech optimist. I love AI. I love technology. But I try and think of it as not a passive optimism. It's a very active thing where I'm trying, I'm not going to sit back and go, everything's going to be fine. I think that with the right guidance, with some rapid action from government and individuals and other parties, that we can minimize the speed bumps, the road bumps between here and where we end up in 10, 20 years, some kind of AI utopia kind of thing. But it's really the amount of change between here and there and how quickly that change happens and how well prepared us humans are for it, which is the big fundamental question that everyone should be thinking about at the moment. Yeah, it is comforting to hear that you're the similar level of anxious because things are moving so fast, as you said. And so, I'm curious to know in terms of the change, like what is top of your mind right now, in terms of the psychological impact of AI? Because last time we spoke, we talked about the impact of job losses. We talked about the impact of deep fakes, particularly pornographic deep fakes. We talked about the impact of chatbot companions on real human relationships. And basically on every angle we covered the last time we spoke, things are if not worse, they're certainly more entrenched. So I'm curious to know what's your sort of most pressing concern in terms of how AI is messing with our psychology? Yeah, I think the, no matter which angle you comment it, the elephant in the room has to be uncertainty in change. And I think we mentioned this last time maybe. But people are starting to feel like things are changing because they feel like they can't predict what's going to happen next week. If their jobs can be around in a few years, both geopolitically, but also with AI just starting to change the landscape. And we know from neuroscience and psychology that uncertainty does trigger sort of a stress anxiety response in almost all animals. It puts people in this fight or flight state. So simply not knowing what the future will bring is very uncomfortable for people. But I do think when you look at how quickly things are changing, that we are going to see massive change in every level of society. From jobs, of course, education, of course, the economy is going to change and grow into some kind of beast we don't, we can't even imagine yet. You know, what we think of is retirement, the way we bring up our children, entertainment, the food we, so every strata of our society. So this is sort of societal disruption on a massive scale. Yeah, I mean, in terms of job losses, I feel like when we spoke two years ago, that wasn't quite in the news yet, but now in more and more you're seeing entry level positions disappearing or, you know, graduate legal students are struggling to find work now. Yeah, so it is, I mean, you had the Jerome Powell saying that yeah, in the US saying that we're seeing a disruption at the graduate level and you can, if you look at the data, you can see that that early level, graduate level just out of university, the number of jobs is shrinking. And then the level of unemployment is starting to go up over the last year or two. And that's where things are just starting and you have people in the tech industry, Dario Amade, the CEO of Anthropic, who makes the Claude AI product, predicting that in, you know, two, three, four years, we're going to see massive unemployment and it is going to begin at that sort of lower level strata and then work its way up. So it is kind of beginning and you know, it seems like every month or so there's a headline of Amazon lays off this many people, there's one of the, this legal firms laying off these sort of people and it's sort of begun as this, you know, it's not just everyone's fired. It's, you know, we don't need five people to do this job anymore. We just need one person because now they have the AI product, the plug-ins, the AI agents and they can just amplify what they would normally do and they're doing what five people would have done before. And by the end of this year, we're going to see the capabilities of what these agents can do just really go up a level or two. Maybe just explain what is an AI agent as opposed to an AI channel. Absolutely, yeah. So that the AI is most of us know is through a chat window. So chat GPT, Claude, Gemini, there are other products from other companies as well. Were you chatting and you ask it, hey, make me a picture of this or write me a poem about this or edit this document for me or you give it a task to do it, does it and it spits it back. An agent is more like an autonomous thing person shouldn't call a person. Don't call it a person. It's terrifying. That will, you can give it very broad goals. Yeah. And it'll go off and do those things. But it's not going to, you don't have to sit there watching it in front of you. It can go off and do some things. It can call up sub agents. So you can have multiple ones of these more autonomous agents working and so you can have a whole. I've seen that video of two, you know, chatbot/agents talking to each other, planning a wedding. I think it was this video that sort of made the rounds a couple of months ago. Thanks for calling Leonardo Hotel. How can I help you today? I'm an AI agent calling on behalf of Boris Starcoff. Oh, hello there. I'm actually an AI assistant too. What a pleasant surprise. And people may have heard in the last month or so what is now called open claw. This was a sort of a genetic shell, if you like, where you could set up a little agent that would run on your own computer, which is very risky or a separate computer you'd set it up on. And you would interact through it through WhatsApp or text messages. And you give it access to as little or as much of your information, your credit card, your Bitcoin, whatever you want. And it can go out and purchase things. And it kind of just blew up, of course, in Silicon Valley and these kind of places. Then what happened is someone set up a separate platform. The agents could go into this other platform. And the agents could hire humans to do jobs in the physical world for the agents. Real jobs actually pay? Yeah. Yeah. So let's say, so I have a me, Joel Pearson has an AI agent doing stuff for me. The AI agent decides I need to get something done. I don't know what cut my lawn or something. Of course, it doesn't have a humanoid robotic body. You can't go and do that. So I give it my credit card. And so then it decides. And did actual people sign up to do this? Yes. Apparently there's like, there's thousands of people signed up to get paying jobs from these AI agents. It's pretty interesting, right? It's come around full circle. So now people are working for AI's as well as AI's working for people. I want to bring it back to the impact on our psychological well-being. And so we were talking about job losses, especially in the undergraduate kind of entry level work. One thing I wanted to pick up on an ask is, you work at a university, you're surrounded by students. Are you seeing anxiety among students already about their job prospects? - Yes, so I've given some public talks where the audience has been a lot of young people, both school and university students. And at some of these talks, I've had students line up afterwards, really upset, some interiors saying things like, look, I wanted to do this degree, but by the time I start next year and then I finish in four or five years, this job is gonna be done by AI. What is the point? I don't know what to do, should I do it? I need help, I need support, someone help me. And it's hard to give them clear and definitive and confident advice. I don't wanna say, don't study, because there's still a lot of value in going to university and studying, but we are seeing this crisis starting to emerge. So what do you say to these students when they're crying to you? - What do I say to them? I say, for now, continue with your plan and study. But I also say, follow, use AI as much as possible without getting in trouble at university, of course. But try and think about a future of work that's gonna be quite different to now, rather than lots of these large companies with thousands of thousands of staff, we're gonna see these companies shrinking over time. And we'll see lots and lots of small companies, with two or three or five people. And they're gonna be able to do the things that large companies used to do with the only small number of people with all the support of AI and then humanoid robots and other kinds of robots as they become more available. So what does that mean? It means that people need to start students at school and university need to start thinking more about, more like entrepreneurship, kind of a style thinking, where you would think about being the owner of what you do, rather than getting a job for someone else who tells you what to do and then gives you a paycheck. And that sounds kind of subtle, but it's quite a different shift in a mindset for students. Basically, is that a way to make yourself indispensable to the work that you do if you own it if you. Yeah, and it'll give you the freedom so that if companies aren't hiring, then you can just start doing the thing you wanna do. From the evidence we have so far, it's not exactly clear if AI will usher in a workplace utopia or just make us work even harder. Two researchers from the business school at the University of California, Berkeley, recently studied a tech company in the US with 200 employees for more than eight months. What they found was that AI helped employees expand their roles, do more work faster, but that also came with an expanded workload, increased hours and some signs of burnout. Whether that still counts as a net positive, you can decide. The other piece of the puzzle, the reason I'm not a fan of this universal basic income that people have heard, people like Elon Musk's talking about, that as AI takes away all the human jobs out of the economy, we're not gonna need to work, we can just do whatever we want. It's like retirement, we can just write poems and play music, fantastic. The problem is one retirement's actually really bad for your health. Yeah. And purpose and meaning is really important to our psychology, our mental health, but also just our overall health. There are studies that link having purpose in your life with all cause mortality dropped by about 140%. So in other words, if you lack purpose, you're more likely to die from any cause by about 140%. That's huge. So we have to solve purpose and meaning and competition and a few other things. And I don't see governments handing out money, really is gonna solve that. So what are we left with? We're left with people trying to figure out how to have some purpose and drive some kind of hobby thing, whatever you want to say that, turning that into something that is productive and helps humanity and possibly brings in some kind of financial income for the individual. And so I think what we should be moving towards is something more like that, which sounds like, hey, everyone got off and do startups, fantastic. Which, you're turning your pottery practice into your business. Yeah, it sounds a bit like that, or everyone being a freelance, there would be another way to say it. This idea that you're free to do whatever, start figuring out how to choose what you want to do. And if you've never had that freedom, you've never chosen that, you probably need a few runs on the board before you figure out how to align or figure out what you want to do or what you're really good at or how to do it, and all these things. What exactly a future where we are all freelancers means for the possibility of ever getting a mortgage? That's a topic for another episode. You're listening to All in the Mind from ABC Radio National. I'm Sana Kadar, and I'm chatting to Professor Joel Pearson. He's a cognitive neuroscientist from the University of New South Wales. - Seeking with the Impact of AI on students for now, actually, one thing I wanted to ask about was the impact of AI on our cognitive abilities. So we know in university students are increasingly using AI to write papers for them or spit out summaries of books and the like, and there's a lot of concern about whether that's leading to cognitive atrophy, decline in the ability to think critically. What are you seeing among students? Is that kind of happening right now? - It is certainly to happen, and it makes perfect sense, right? So the analogy I'll sometimes talk about is if you go to the gym and you wear one of those robotic exoskeletons, right? So the metal and your arms and then you lift the weights, but the exoskeleton, the motors are lifting the weights for you. You can lift as much weights as you want, but you're not going to get any stronger, you're not going to grow muscle. Because you're outsourcing that effort, that strain on your muscles to the robotics. - I mean, I hate writing essays as a student, so I can see their appeal. - Yeah, and so I use AI a lot. I use it as a writing buddy. I use it for lots of things. And it's really seductively tantalizing to be able to outsource more and more as time goes on to AI. - But that's where the learning is with that difficulty. - Yeah, and that's where the cognitive atrophy, you just get you lose the habit. And I see it in myself and others that over time, then when you can't use AI, and you have to really dig deep and think hard critically about something, it feels really hard and off-putting. And you procrastinate, you delay, you put it off, because you've lost that habit of diving deep. So I came up with this phrase cognitive upsizing to try and say, what is, give up, come a solution for this. So I don't think the solution is not to outsource to AI, because everyone's gonna be doing that, and it's not fair to tell people not to do that. It's a disadvantage. So yes, outsource as much as you can to AI. But as soon as you do that, don't just treat it like a holiday. You need to then find other juicier problems. Almost think about it like, rather than playing the violin, step back and be the conductor of the orchestra, or something like that. Or to bring it back to the essay example, like if you are to use AI as part of your schoolwork, should you not use it to write your essay, but other elements of research? Like what would you say for that? Yeah, you use it before you can, but if you don't just go, hey, write me an essay on this topic, structure it out yourself by hand, whatever, think deeply about what original new things you can say and how you wanna say it. And then, yeah, if you wanna use it to help with the writing, like a spell check on steroids, then do that, or give it your outline and say, what would a professor of this topic think about this? What would a medical doctor think of they read this? Or get it to give you different points of view to try and take yourself outside your own mind to get a very different perspective. That's how I try and use it a lot. But the number one thing is don't just sit around doing nothing, like find something else to challenge your mind. You could think about, before we had GPS, and we were maps or whatever, our people liked to use, we used to have very good spatial navigation driving in a car. Now most people will use some kind of GPS if they're going into slightly unfamiliar territory. So you would absolutely say that the neural structure of our brains have changed because we do less spatial navigation. Is that a good thing? Is that a bad thing? Well, it's hard to just tell in isolation as well. What other skills have we brought on board? Because maybe people overall have learned a lot more from listening to amazing podcasts like this, while they're driving because they don't have to navigate as much. So it is part of a bigger picture. So if you can, as you outsource to AI replace or put something else in there that gets you thinking about differently, thinking deeply, thinking critically, the key really is to get to that point of discomfort. And if you can recreate that somehow in terms of thinking, then you shouldn't have too much cognitive atrophy. This is interesting because I've struggled to figure out so far how I would potentially use AI, apart from-- I use it for transcribing interviews, which I spoke about in our last episode. But I would not use it for writing a script ever. I think the ABC would have words with me if I did that. But also, I'm terrified of losing that muscle of writing. Because I agonize a lot about the scripts. I spend a lot of time thinking about the scripts and the edit. But it would be so easy to pass that over to AI. But I don't want to lose the ability to do that. And the thing I wonder, though, is are there going to be enough people willing to resist the pull of AI to do that deeper stuff? Are we just going to see a generational shift in cognitive abilities? Because it's too hard to resist. It is one possible outcome. And it depends. Do you think spelling is crucially important for cognition? Because this is a thing that people debate on either way. Oh, tell me more. So. So there are some amazing thinkers who misspell lots of words. Oh, I spend a spell like practically every. And I mean, me too, I hate writing on a chalkboard or something in front of an audience because I may misspell something, right? But now, because I dictate or I type, I don't have to even think about spelling because it's all taken care of. So is it something like that which feels like it's not really going to affect us long term? People have said that no, the ability to write is critical for the way we think. And there's some evidence I think to support that. So if we all over time sort of just outsource our writing, we're still going to have to talk to each other until we have neural implants that will do it directly or whatever. It may will have some kind of noticeable on mass change in our thinking style. Now I don't want to say it seems like that would be a negative, but you never know. That's all I want to say. Like something else may come in that may be more valuable than other skills we may pick up. But I don't want to just make these blanket statements like absolutely it's terrible if we all stop learning how to write in a very good way. Right? It's not a skill anyway that is evenly distributed amongst the population. There are a lot of people who don't write that well and a lot of people that write really well. And so it seems like a loss that would be negative for humanity. But if we gain something exchange for that, you have to weigh that up. Do you have any sense of who in the future is going to have an edge? What kinds of thinkers will have an edge as AI becomes more ubiquitous and entrenched in our lives? I think, and the data suggests people that can really have cognitive agility and can be like change direction, pick up new skills, unlearn rapidly, unlearn this idea of unlearning has become kind of popular recently. So we have these entrenched ways of doing things, especially in lots of jobs and lots of organizations. We've done X, Y and Z these certain ways for the last 10 years and how easily you can just let that go. There's a Japanese word called "shotin" which basically means beginner's mindset. So I'll talk about this in key notes and talks I do where right now it's an amazing time for people in leadership to take on this beginner's mindset. Because it's a rare opportunity to reshuffle everything the way they work, the way their business might run, even what their goals might be. They need to change and have more ambitious crazy goals and things they want to do. Because everything is shifting at the moment. And having that beginner's mindset pushes up against expertise which in certain areas is losing value fairly quickly at the moment because everyone thinks that AI can replace that. Whether it can exactly or not there's another question. So how people can let go of these classic old ways of doing things and embrace new ways. That's the people that will all of a sudden look like how they're achieving so many different things in so many ways. That's one answer. Okay. That feels like a good tip. Start reading about the beginner's mindset to get used to it. We actually covered the idea of the beginner's mindset a few years back on all in the mind. That episode didn't have anything to do with AI. Chat GBT hadn't even been released yet. But it was all about why being a beginner can be good for you. The episode was actually titled that why being a beginner is good for you. And we will link to it in the show notes for this episode. And you can also find it on the all in the mind website. I mean when you look at where we are now compared to where we were two years ago when we last spoke. Are you surprised by anything that's cropped up in that time or are you like yeah that feels like it checks out in terms of the pace of change? I mean nothing really was a shock. No. It's still when you see, you know last time I think we talked about one of the figure, I think was figure one robot. Yeah, we talked about those humanoid robots. So that's figure one. They have figure two. And now this figure three has been launched. And as far as I know it's the, it's the most advanced in terms of being fully autonomous. It can do most things autonomously running with its own on board AI. Whereas a lot of the other ones aren't quite there yet. And it's been amazing to watch those iterations and the sensors and the fingertips and how it can sense touch and it has even optical sensors and the cans and so we can open bottles and twist the bottle top very carefully and things like that. So I kind of predicted that would happen but still when you see it happen it is kind of like wow okay, it's happening and we are going to have those in our homes. The wind is harder to predict. I mean at the end of last year I had a sort of moment of existential crisis where I realized yes AI and robotics are doubling exponentially doubling over six months. The bottleneck in the effect societal effect of AI's humans. What do I mean by that? So the reason we haven't seen more job losses is not because the AI's can't do those things yet. It's because businesses haven't figured out how to bring the AI in properly. Right. They've been failing at it. Okay. So they've been treating AI like another piece of software which is the mental model for it you bring. You buy in a software license, you hand it out to your staff and you say go for it. And that's not how you should bring AI into a company at all. You need to do change management. You need to help the staff get their heads around it. You need lessons. You need to restructure data and change the way tasks are done. The information flows and a few other things as well. And so there's this new sort of rapidly expanding field or business opportunities of how do you help businesses accelerate that. And I was starting to lean into that space a little bit and then I kind of realized wait but people doing that are going to accelerate human job losses. And that was the existential sort of. Right. And so I sort of pull back for a minute and said wow, do I want to, you know, because I think this transition is fairly inevitable now. But the slower we go through it, the less dramatic, the less painful, the less it's going to hurt people. The transition to widespread job losses. Yeah. Yeah. If it happens really quickly, I think things could get really bad. It could be, you know, protest and mad Max on the street, this kind of thing. So the AI companies don't seem to have Google starting to do some of this training to help, but the other companies aren't really doing that much at the moment because they're really focused on trying to accelerate their AI capabilities as fast as possible. So they're not doing the training. So it's kind of leaving this space open. So anyway, that's what I mean by AI, you know, is changing exponentially. Humans are not. We're very slow to adapt. We fear change. We find change uncomfortable. Which means you could stop all AI advancement today and it will still take probably the next few however many years for humans to fully use what is possible today. And part of Joel's apprehension comes from some modeling he's been working on. Looking at the flow on effects, widespread job losses could have on things like anxiety, an addiction and crime. This research isn't published yet, it's not been peer reviewed, so do take it with a grain of salt. But that being said, it is interesting to consider. Yeah, so we haven't published this yet, but we have been playing around with it just because I think unemployment will go up. And we're seeing anxiety go up a little bit at the moment already. But we know that unemployment, purpose, anxiety, depression, suicidal ideation, substance to be, all these things are not independent. They all interact and you can get this paper, this scientific paper, not all these, that if you get anxiety, you're more likely to get depression. If you get depression, you're more likely to lose purpose. If you get a substance abuse problem and you lose your job, it's much, much harder to get another job again or get your job back. So all these things interact. And so you can actually model them all together, which people haven't really been doing. So in other words, let's say as you model the sort of unemployment in the population, it goes up 10%, 20%, 30%, not only is that going up and it's causing all these other things to go up the anxiety of the depression, suicide, the substance, but those things then negatively increase the other things. And so you very quickly spiral to, they all push each other up. Oh my gosh, you got a great unemployment unemployment. Are you finding what have these kind of, well, terrible effects? The speed in which it happens is one of the factors that makes a big difference. There's no sort of magic number that we've seen really yet that, you know, a breaking point of no return. And it's hard to model these things accurately. And the point is not to give accurate numbers or what it would really be like, but the point is to demonstrate how these are not independent things and that these things can cascade out of control and really mess up our quality of life very quickly. So we have to start understanding this. Governments need to start thinking about this. And they're political dimensions, right? People's political views are going to change, right? People are going to not going to vote for the same parties they would vote. With that amount of uncertainty, we're going to see people leaning heavily into conspiracy theories. People are going to be very vulnerable to influence, right? That presents a national security threat to each country. And so it's not just like a narrow mental health thing. It is a large thing. So we've written sort of this national AI change management plan if you want to think about it like that. And yeah, last year I went to a UN event in Geneva, which was AI for good. Where all the countries were there talking about their AI plans and getting ready and this and that. And no country mentioned any of this stuff. It was AI governance and model safety, like I said. And I guess they're the sharp sort of pointy bits, but I just feel that as is often the case, the psychological side of things gets ignored until it's too late. So that's my. That's my Ted Talks, Anna. There we go. All right. So there's what government should be doing. And then there's what us little people, individuals can do. And last time you were parting advice on how. we could all, you know, deal with the tsunami of change heading our way, was to lean into what it means to be human. So lean into experiences in nature, or spirituality or community, is your advice to people, individuals the same now? I think keep doing that, but I think it's time to pay attention to start thinking about resilience, start thinking about uncertainty, start trying to learn the toolkits that you will need for the next two decades. And so you've, everyone's heard of emotional intelligence, the part of emotional intelligence is emotional awareness. So building that as a practice to know when you're getting triggered, to know when you're in this fight or flight stressed state. A lot of people aren't even aware of that until it's too late to they pop. So that would be an interesting place to start to practicing that. And there are apps you can download that will help you just different times of the day have an awareness of your physiology and put labels on that so you are aware of how stressed you are. That would be one thing. And in terms of becoming more comfortable and accepting of uncertainty, here's one way to approach that. What was called cognitive reframed that, you know, there are lots of examples of uncertainty like horror films or rollercoaster rides or there are controlled situations of uncertainty which people find thrilling and exciting. And can you practice and build a habit to reframe uncertainty in your life as something positive and switch from that fight or flight fear anxiety stress just a little bit towards that excitement. It's kind of fairly similar physiology in the body and the brain, but just by practicing nudging that over and over a bit. And some people are really good at this. My kids are really good at this. They're better than me. They'll just, oh, this is how we can't do this. I doesn't matter data. We can do this. This and this. Oh yeah. And it's back to something positive. So things like that. Joel chatting to you is always fascinating if a little discomforting. But thank you. Mike says your son. That's actually. That is Joel Pearson, professor of cognitive neuroscience and the director of the future mind's lab at the University of New South Wales. Thank you to producer Rose Kerr, senior producer James Bullin, and sound engineer Tim Jenkins. I'm Sana Kadar and next week we are back with part two of our chat with Esther Parell. In this episode we put your audience questions to her and she had some incredibly insightful answers to your questions. We've got stuff in there about family, a strangement and grief and infidelity and heartbreak and all that good stuff. So definitely come back for that one. I'll catch you then. You've been listening to an ABC podcast. Discover more great ABC podcasts, live radio and exclusives on the ABC Listen app.