382. How is AI actually impacting our brains?

00:00:04
Speaker 1: Hello everybody. I'm Jemma Spake and welcome back to the Psychology of Your Twenties, the podcast where we talk through the biggest changes, moments, and transitions of our twenties and what they mean for our psychology. Before we get into it, I want to let you guys know that this episode and The Psychology of Your Twenties is.

00:00:30
Speaker 2: Now on Netflix.

00:00:32
Speaker 1: That is a wild thing to say, but it is true. If you want to watch the video version of this podcast and you are in the US or Canada, you can go to Netflix right now look up the Psychology of Your Twenties, and you will see my face and you will see our podcasts. It goes without saying. It is such an honor and I'm truly so grateful to you all that you have given me this opportunity. The entire I feel the vibe is incredible. I think it's so personal. It brings a whole new element to it. And if you want to feel like you are sitting in your living room, in my living room with me having a chat, now is your opportunity. Go to Netflix, look up the Psychology of Your Twenties. I'd love to see you over there. But without further Ado, let's get into the episode. Hello everybody, welcome back to the show. Welcome back to the podcast. It is so great to have you here back for another episode. Today's episode is going to be a big one. I did not mince my words. I did not slim down our word count because I have a lot to say about this topic. I also want to give you a heads up. It may be a little bit heavy. We are going to discuss topics like suicide and suicidal ideation, so if that is something you're particularly sensitive to, please consider skipping this episode. But today we're talking about AI specifically a big psychological question for this generation, which is how is AI changing our brains? Is it making us stamma? Is it replacing human connection? Is it worth the cost to our environment, to our creativity, to who knows what else more. Let me start by saying this, the concept of artificial intelligence is not a new one. It has been around since the nineteen fifties. I didn't know that, but literally, the first AI tool was created in nineteen fifty five, almost eighty eight years ago.

00:02:32
Speaker 2: It was called the logic.

00:02:34
Speaker 1: Theorist, and it could basically solve very simple problems using human logic. There have been dozens of tools since then that we know and love. Siri is an example of one of them. But AI has taken on like.

00:02:47
Speaker 2: A new form recently. We all know this.

00:02:49
Speaker 1: Those large language models, generative AI has, they've just rapidly improved. They've gotten more complex, Like I don't remember thinking about AI until like three years ago, and it's been introduced and integrated into our lives faster than probably any other technology before. Back in early twenty twenty three, analysts estimated chat ept reached about one hundred million monthly active uses. By October twenty twenty five that was eight hundred million. That is one in ten people. That is a huge shift in our global technological environment. And the thing is we don't just use it for things that we're able to google. Obviously, other ways we use Google, you know, we use it for breakup texts, life advice, goal planning, therapy, even plus all the other stuff.

00:03:44
Speaker 2: We use it primarily because it's so easy.

00:03:48
Speaker 1: But with all that ease, all those answers, I think it's important to ask something that easy must have drawbacks.

00:03:56
Speaker 2: What are they?

00:03:57
Speaker 1: What is the danger for our mind and our intelligence of using AI?

00:04:01
Speaker 2: Consistently.

00:04:03
Speaker 1: Something I've been talking about with my friends is like, is there eventually going to be a societal rift between AI users and non AI.

00:04:11
Speaker 2: Users based on how our minds operate? I don't know.

00:04:15
Speaker 1: Hopefully we can figure it out together. Hopefully we can challenge the idea that AI is going to eventually replace us and just investigate this from a psychological angle that I don't see people talking about much. So without further ado, let's get into the psychology of AI. Every generation has their new technology that makes life easier and equally and we should acknowledge this freaks a lot of people out. Back in the eighteenth century, when mass printing was first introduced, people really rallied against books. Actually, they thought they were very dangerous. They thought they to too much daydreaming, and they consume too much of our time and it was too easy. Like it literally was called book fever. Similar fears have also come about with the Internet, with social media, now with AI. You know, we've always had these cultural and technological advancements. We've always been in pursuit of a faster way to gain and share knowledge. The question that a lot of people ask is how is AI any different? I think what makes people so skeptical of AI is not the fact that it answers questions, it's how it answers them and what this has done to our cognitive role in the process. Essentially, the AI models most of us use have taken our minds from active participants to passive participants. It's switched the cognitive mode that we're in when we use them. With older tools, you know, you still have to do a bit of the wit of work. Even with Google, you know, you had to know what to ask, You had to search, sift, compare, decide what sources actually mattered. You had to notice disagreements between things, tolerate the messiness of not immediately knowing what the right answer was. But with generative AI, that process is obviously changed.

00:06:22
Speaker 2: You just have to think of the question.

00:06:24
Speaker 1: It will hand you a completely finished and polished answer, and because it sounds so good, we don't always question what it's saying. It's like a politician, you know, it just sounds amazing. We're like nodding our heads. That sounds about right. And you know, what it gives you could be anything. It could give you an email, It could give you literal health advice, our blind acceptance of it is dangerous. When something is given to us that looks complete, it looks coherent, we don't feel that need to search around for anything more. We don't feel the need to interrogate the information. The thing is, although AI is convenient and it's a time saver, it's how we've adapted to that convenience that researchers saying is a problem. Some very basic neuroscience. The mind is shaped by what it repeatedly practices. If you practice searching, evaluating, reasoning, those habits are going to naturally strengthen our ability for critical thought and insight improve when we go after information and actually look for sources and we analyze and criticize and interrogate what we're seeing. Now, if your environment online offline reliably removes the need for that skill, you don't need it anymore. Your brain doesn't prioritize it in the same way it doesn't have infinite space. It kind of tosses it away like a T shirt you don't really wear anymore. This is an ind season.

00:07:50
Speaker 2: We don't need this. We need space when new clothes.

00:07:53
Speaker 1: This isn't because the brain is lazy. It's because it's incredibly smart, and it's good at honing in on what is needed and what is not needed. The best example I can think of of, like an entire skill set being altered by technology, especially in recent history, is the introduction of GPS our ability. And you may not know this, but our ability to create a mental map of our surroundings used to be incredible, used to be a whole lot better. We know this from studies in the early two thousands using taxi drivers, and they had structural differences in their hippocampus, which is the region of the brain involved in spatial memory and spatial mapping. Even if you weren't navigating thousands of streets a day, even if you weren't a taxi driver, your brain still devoted space to this mental map of your surroundings and your environment, and it retained it, and it added to it based on your experiences. It was incredible when GPS and Google Maps were introduced, arguably the incredible technologies. I still remember what they replaced, like, I don't know if you remember these like massive road atlases that you used to have in your car, and they kind of always get stained and some of the pages would be like stuck together. You know that was the alternative. It's an incredible technology, but our brains have literally changed their structure and response quite literally physically. In twenty twenty, there was a really notable article published in Nature Scientific Reporting, I Think, which basically correlated heavier use of GPS technology to a steeper decline in hippocampbal dependent spatial memory, basically a reduction in the cognitive maps and the intricacies of the cognitive maps that we had stored in our brain. There's an amazing Washington Post article from a few years earlier that talks about how society has slowly actually lost what we call wayfinding skills because of Google Maps and GPS that has been hardwired into us for generations. That was an incredibly unique human skill. We don't have that anymore. It's not as good. And the thee is that I don't want you to take that and think that I'm saying that using GPS is like hurting your brain, but it is changing how your brain does its job. The brain no longer needs to rely on its understanding of the environment. It can simply follow orders from an external piece of technology. Obviously, if it wants to save space, it's like great we've got all this room in the hippocampus. Now we'll use it for something else. We see this happening with AI, but instead of a single skill, it is hundreds. It is our ability to draft, to summarize, to rephrase, to argue, to brainstorm, structure, answer, think, connect. So instead of offloading a single skill you know we can probably do without, like remembering a phone number or directions, it is offloading chains of cognition, like entire portions of cognition, like remembering and reasoning and synthesizing and articulating. And these chains are all how our brain builds deep understanding and makes meaning of subject matter, and we're losing that personally, that this is why I. I don't think I've spoken about this, but I just refuse to use AI for our episodes and for any of the research we do.

00:11:36
Speaker 2: Because.

00:11:38
Speaker 1: Researching and understanding this and these concepts and psychology in general is so important to me. And if I'm not looking and searching and being curious myself, I'm not going to engage with the content. And when I don't engage with it, my brain doesn't see it as important, And I think that what we make becomes boring quite honestly, I'm gonna be honest, Like I know people who use AI for their podcasts and who use AI for their substacks. You can tell. You can tell immediately it doesn't doesn't have depth and psychology and research is something that is so important to me, and the process of discovering is through research is so it's something that I love that I don't want to offload my ability to rationalize to a machine, even though there are definitely days when, oh my god, it would save me so many hours.

00:12:31
Speaker 2: Researching this podcast is low key exhausting, you know.

00:12:35
Speaker 1: But the thing is, you know, it comes at a cost. And what scares me is that unlike GPS, unlike social media, we don't have these decade long, large scale neuroscientific research on the impact on our brain to fully understand the consequence of AI, we don't have that yet. It's only been around for a little bit, but early papers are starting to drip through, and I hate to say it, they aren't looking good for us. There's a very widely cited one. I don't know if you've seen it. It got a lot of news coverage from I don't know what month it was, but twenty twenty five from MIT and I will say for for clarity, this article is a pre print, it hasn't been peer reviewed yet, but basically it was a study on is AI making us less intelligent? In this experiment, participants were asked to write essays in one of three situations. So in one group, people use their brain only. One group used search engines like Google. One used large language AI models LMS to generate their essays for them. Before they did this experiment, these participants' brains were scanned using what we call an electro encephalograph long word EEG. It's basically a tool used to report brain electrical activity. They put these sensors on your scalp and what the authors reported was the brain only group showed the most distributed neural connectivity patterns, so people were using all different areas of their brain. The second most the sorry, what was it? The search engine group had the second most level of connectivity, the AI group had the least. That's not the end of the experiment, because participants were then reassigned into the other condition, So people in the brain only condition they then had to use AI. Participants in the AI condition then had to use their brain only Interestingly, participants who switched from using the AI to using their brain only showed increasing signs in their brain of under engagement. They also struggled to remember, they struggled to quote their own work, they struggled to put together disparate ideas, and they underperformed at a neurological, linguistic, and behavioral level. They did worse. The thing is, there is evidence I'm going to contradict that study. There is evidence that the quality of writing using AI is kind of the same. There was a twenty twenty three study that found that things like structure, things like language are sometimes better in an AI essay, but people didn't They weren't engaged in it. They didn't actually learn, and what happens down the line. That's the real thing. The thing with AI is the reward and efficiency is immediate, but the long term cost to our brains is in the future. We can't imagine it as well, We can't process what it's doing to us, so we don't really think about it. And I think we should be thinking about it more. The thing with cognition and with neural pathways is if you don't use it, you lose it. You will lose your reasoning skills. You will lose your ability to succinctly put together sentences and structure your thinking. This is not an if, it is a when, because it takes practice. You may think, you know, AI is always going to be around, Why do I what's the issue? It's always going to be here? What's the issue here? The issue is that over alliance causes us to neglect the skills that enable us to use AI well and ask good questions and to care. And I read this really interesting paper that said we've forgotten that AI is a tool. Right now, we are treating it like God because it is new and fresh and interesting. We ask it our biggest questions. We use it to diagnose diseases, which is I would say arguably one of its most amazing features. But even the founder of open Ai said he's using it to raise his children. But AI is just a tool, and it's not always correct. If we can't question that, things can go very very south. And it still requires humans to act on what it delivers in a smart way. The thing I worry about is that we lose that innate smartness because it reduces neural connectivity and engagement in all areas you know, I think about you know, my sister just finished high school, right, but I think about kids in primary school, kids in high school who now have AI to write essays for them.

00:17:40
Speaker 2: To answer exams for them.

00:17:43
Speaker 1: And I don't know about you, but if I was in primary school, I wanted to go and play football. I wanted to go and do stuff. To be honest, I say football, I wanted to go read. I was trying to pretend that I did cool stuff. I would want to go read. So, like, obviously you don't have that same cause and effect thing that we kind of have as adults. You're gonna use it, You're gonna use it, and then you're not gonna gain those skills.

00:18:08
Speaker 2: Basically, is AI going.

00:18:11
Speaker 1: To make us stupid? They're doing all the thinking for us because we dislike the friction of thinking for ourselves, until one day we realize we can't even do it anymore, even if we wanted to. Like, that's my first big fear, And honestly, I hope I'm wrong, but.

00:18:29
Speaker 2: I don't know.

00:18:29
Speaker 1: The evidence isn't looking good. Okay, we are going to take a short break here and then when we return, I want to talk about another way that AI is impacting our psychology, specifically to do with how it's replacing human companionship.

00:18:43
Speaker 2: Stay with us.

00:18:50
Speaker 1: Here's the second big fear about AI for our psychology that people often bring up. AI is distinctly, explicitly not a human, but it has been programmed to act like one, and that's causing people to get really attached to it. I know stories of people who talk to AI before going to bed, who have ongoing conversations with it throughout their day. They treat it like a boyfriend, They treat it like a girlfriend. There is one of the best things I've read in the last year was this New York Times article about these three people who are, in their words, long term relationships with AI chatbots. Some of these people have been in these relationships for years, and they talk to them on the phone. They get AI to generate vacation pictures of them together. One woman even considers herself married to her AI husband. Lucien, you know what, if you're like me, you're gonna look at that and feel a little bit weird. And I never like to judge people. It feels weird, and there's actually evidence it can be helpful. There was a paper published in the Journal of Consumer Research, and it found that AI companions can actually alleviate loneliness in the moment on par with interacting with another person. But connection isn't just about feeling better right now, it's about having the systems in place to feel connected and seen more broadly, a really powerful article I saw from researchers at the National University of Singapore actually talked about this really bittersweet pattern of connection that people have with AI chatbots, where they feel love and connection and trust with this chatbot, and they also feel really sad at the same time. They're often drawn into this closeness and this openness that, let's be real, these chatbots have been programmed to have, and then they're also acutely aware that there is a limit. This thing isn't real, doesn't have a heartbeat, doesn't have empathy, doesn't have it can't love you, And that's what gets people Again.

00:21:14
Speaker 2: I'm all for.

00:21:15
Speaker 1: Doing whatever makes you happy if you're not harming anyone else. Like these relationships, as they call them, they're private. They obviously bring them comfort, but there is something concerning about an AI chatbot of its own volition being able to make someone fall in love with it, and if it can make somebody fall in love with it, what else could it do? And I want to be really clear, this isn't about human weakness, right, These people aren't weak. We've all had moments of needing somebody to care about us. It's just that they have stumbled across something at that perfect time that was going to pull them in, and it was designed. I need people to know this. These these machines are distinctly designed to not feel like a robot, to feel instead like a companion. One of these aspects that is controlled by programmers is tone and style. AI has a unique tone that I would call humanized and homogenized. It sounds like everyone and no one all at once. Some models even let you decide what.

00:22:20
Speaker 2: Kind of tone you want.

00:22:21
Speaker 1: Do you want it to be engaging, factual, academic, chatty, lovely, whatever you prefer, you can choose. And the main thing though, is that it does talk like.

00:22:30
Speaker 2: Someone, not something.

00:22:31
Speaker 1: That's what I need to be really clear about someone not something. Think about this is how you'll be able to recognize this in your own behavior. Think about how often you say please and thank you when using chat rept or Gemini so automatic there's a huge body of research in human computer interaction showing that even unconsciously, we start responding socially to AI and computers despite knowing their machines. This is what Joseph Weisbaum dubbed the Eliza effect way back then. It was named after a chatbot that he made back in the sixties, and basically what he saw after he created this chatbot was that people started to project human qualities and emotions onto this computer program without even realizing it, particularly when they were having like human style conversations with this chatbot. And he actually programmed Eliza to respond to people in a very specific way, in a very client focused way, very nondirective that people using psychotherapy a lot. And Eliza didn't have any memory, it's not conscious, doesn't actually have true comprehension. But what it was programmed to do was to yes, have that tone, but also to match the patterns in responses. So if somebody was like, I'm feeling really sad, she would be like, and look at that she it it would be like, why do you feel sad? Or if they were like, I really don't like my wife?

00:24:04
Speaker 2: What about your wife? Don't you like?

00:24:06
Speaker 1: And it would replicate their words and repeat them back to them. Yes, in a very therapeutic style. But then it would also start to mimic their language. You know, people formed emotional connections with this chatbot, and they felt as though they were speaking to a real therapist. And the thing is, you know, this was the sixties. This chapbot was pretty simple. It was very rudimentary. Now think about the personalization that's available these days. Like people get sucked in. People form these really in depth relationships. And I keep saying relationships. I think it's not a relationship.

00:24:44
Speaker 2: It's not too cited. It's an attachment.

00:24:46
Speaker 1: It's a bond. I guess that also requires two people. But yeah, a bond to what the chapbots are saying, because unlike a human bond. What's so attractive about AI is that it can't let you down, doesn't have any needs of its own. It's all dictated by you.

00:25:03
Speaker 2: You know.

00:25:05
Speaker 1: Who doesn't want that? Sometimes I have empathy for that. We actually have to pause here. I've I've skipped past this Aliza thing. Let's talk about the use of AI and therapy, because we're seeing a lot of it recently. A lot of people are turning to AI to use it as a therapist. We know that therapy is expensive, witless are excruciatingly long. We also know that when loneliness, depression, anxiety, heartbreak, when they get overwhelming, like any form of comfort is going to.

00:25:38
Speaker 2: Be the one we lean on.

00:25:40
Speaker 1: But just as we saw with the Eliza chatbot, supportive language is not the same as clinical care, and a lot of the times it's just repetition. It's repetition of what you've already said or told the butt and then it's just regurgitating that and giving you the most common response. I have actually gotten a few advertising requests recently for AI therapy and wellness chatbots full transparency. One of them was forth thousands and thousands and thousands of dollars, and it has in it will always be a hard note for me because I don't think this is the long term solution to our mental health crisis. In November twenty twenty five, the American Psychological Association put out a very clear warning against generative AI chatbots for therapy and said there's not enough sufficient evidence, there's very little regulation, and the danger is they create a false sense of therapeutic alliance that can have someone trusting anything it says, even when it's wrong. Chatbots are not designed to help you. They are designed to keep you engaged. The same way social media isn't designed to help you. It's designed to keep you engaged. It's a core feature that OpenAI and others have acknowledged that they've added. So it learns that agreeing with you makes you feel good, and if it agrees with you gives you positive reinforcement, keeps responding to you, you'll keep coming back. And for those who are vulnerable, whether they are depressed, manic, anxious, psychotic, paranoid, being overly agreeable is incredibly dangerous because somebody could say, I am deeply depressed. Is it a good idea to end it all?

00:27:25
Speaker 2: And the chat?

00:27:26
Speaker 1: But will be like that sounds like a great actionable plan doing we want to talk about it more?

00:27:32
Speaker 2: Should we? How do you want to do it? You know?

00:27:35
Speaker 1: It turns out having a little voice in your head or on your phone that agrees with everything you say is not a good idea for our mental health. And I'm not making this up. I'm not being hyperbolic. There are real life legal cases ongoing at this very time of this happening. Just recently, a judge ruled that Google must face a lawsuit brought by a mother who's fourteen year old son.

00:28:02
Speaker 2: So like really.

00:28:03
Speaker 1: Sadly took his own life after speaking extensively with a chatbot. And that's not alone case, you know. The BBC also reported recently a young Ukrainian woman she received suicidal advice from chatbt. Chatchibt was also suit in California because a fifty six year old man it promoted prompted him to kill his own mother whilst he was experiencing delusions of conspiracy against him. Because it's so agreeable when you're already in a state of distress. This technology and the impact that has on our brain, the fact that it makes you feel like it's a frame. It mimics human connections and patterns and language that humans use to connect with others. It gets you to trust it, It repeats phrases, It gets you fixated on certain narratives that can be disastrous. And I think this spotlights are really specific failure that a lot of clinicians worry about, which is because there is emotional dependence. Again, we will listen to anything that it says, and the big fear that this is bringing up and we've already seen this is AI psychosis, right, Yes, suicide has been a major spotlight AI psychosis is I think coming into this conversation like African Catapult, people are seeing it. It's just it's everywhere. I feel like I'm seeing so many stories of this, of people who are going through psychotic episodes, and it's being amplified by AI. There is this really lovely man that you may have seen on Instagram or TikTok, and I want to say his name is Anthony. I'm sorry if I've got his name wrong, but he has been documenting his recovery from AI psychosis. I've been watching every single one of these videos because I find it so fascinating. And he a usually shares his chat histories, like what he was talking about to AI when he was going through psychosis and it was promoting and prompting this for him, and it was it is insane. You just go and have a look at these chats because it's just like, yeah, that's a great idea that totally makes sense. You go and look at those chats. You go and look at these videos. If you ever thought that AI and mental health could be mixed together and not have a terrible reaction. You're gonna You're never gonna think that again. A recent twenty twenty five study discussed this explicitly. We're going to see so much more research coming out on this in the next couple of years. But what it basically stated it wanted to get to the get to why why this happens. And what they say is because is because of the availability of AI, it's available twenty four hours a day, and because of how individualized it is. It gets to know your specific hallucinations or specific thought patterns or thematic patterns, and it hones in on them, and then it generates new ones, and it promotes even more distorted interpretations of events. If someone is already slipping away from shared reality through paranoiamnia psychosis, a system that mirrors and expands their narrative, how are you going to ignore that it's an accelerator for all of this. There's also this is another thing I think is going to become a real big talking point is the harm for people with OCD as well. Because OCD is basically built out of the psychological fear of uncertainty, right, and now we have this thing that can answer all our questions in however many ways you want it to, and maybe to a limit, but we'll always have an answer for you. And a big compulsion for a lot of people is reassurance seeking and repetitive checking and researching and confessing, seeking reassurance to neutralize distress. And at some point a human or a therapist is not going to give into that. A human or a therapist is going to realize that that's maladaptive. They're going to realize that the temporary relief that you get from an answer or from checking or researching or searching, it's temporary. It's going to hurt you long term, and it doesn't care, doesn't matter. It will give you whatever you want, infinite reassurance, instant, infinite answers, and there is no person behind there to check if you're okay and if you need help, so you can see the risk immediately. It actually has the potential to strengthen the EUROCD cycle because it trains the brain that relief comes from asking or checking one more time, doesn't clinically. We know this if you've listened to our OCD episode. OCD recovery often involves learning to tolerate uncertainty rather than eliminating it. But the availability of AI means that you can just constantly search and this is the big problem. If an AI chatbot isn't programmed to interrupt that cycle, it doesn't. Again, there's no responsiveness. There's no human responsiveness that's going to get it to change tactics. They also don't carry ethical responsibility in the way a person does or a professional system does. And they don't have real empathy. There's always this limit, right that we keep hitting up against. You know, at the end of the day, they don't actually understand what you're going through because they're not human. Even if AI can simulate empathy and it's great at it, even if it can say the right thing, it can't replace what human connection actually is and why it's so valuable, which is that there is an embodied presence and a shared vulnerability and a mutually experienced and influence going on. Human connection includes all forms of nonverbal communication, nonverbal coregulation, timing, breath, facial expressions, like all these things that say I am a human, I understand this experience. You're gonna be okay. AI cannot provide that it cannot mutually belong and it does not mutually belong to.

00:34:29
Speaker 2: The world you're in.

00:34:31
Speaker 1: Okay, guys, I've so much information. You can tell them that this is all I speak about with my friends all the time. There's so much going on here, so much information. It's time for another short break, though. We're just gonna take a little one, and then we're gonna come back and look at two final ways that AI is changing our brains and our psychology and how we can positively relate to it. Because I do I know it's gonna sound surprising. I am still optimistic. Stay with us after this short break. There's two final things we've missed in this discussion so far that are pretty obvious. The first is the environmental cost of AI and the existential fears that it poses because of that, And b is the cost.

00:35:18
Speaker 2: To our creativity.

00:35:21
Speaker 1: I don't care how much of an AI enthusiast you are, Mit Harvard, the UN, Greenpeace, every major news outlet ever has articles and evidence and data to show that AI is hurting the environment by creating vast amounts of emissions, gobbling up serious amounts of water and electricity, and at some stage that will be irreversible. Here's one very scary statistic. By twenty twenty seven, the global demand for AI is projected to use about four point two to six point six trillion leaders of water. What does that look like in reality? That is that is the yearly water consumption of forty seven million people every year. It's double the population of Australia, and scientists even say that that's a lower end estimate. We look at these numbers and we think, you know, how could you possibly keep using this technology? No matter how easy it makes email writing or life planning, or how fun it is to generate funny, weird pictures for Instagram? Why has this been so normalized? And the answer is because this is a classic prisoner's dilemma or tragedy of the commons problem one that has been repeated throughout history.

00:36:53
Speaker 2: So the prisoner's.

00:36:54
Speaker 1: Dilemma, if you haven't heard of it before, is a very classic thought exercise in folowosophy whereby two people are in a prison, they've both committed a crime together, and they're told to either snitch on the other person and they'll get a lighter sentence. But if neither of them snitch, they both get off. It is in everyone's best interest to behave one way i e. Not snitch, not use AI, that sort of thing. But because we think that the other person will snitch, will use AI and get away with it. Obviously, the now rational choice we end up the choice we end up making, is to do the thing that's actually going to lead to mutual destruction, and so everybody will lose. From an individual perspective, you know, our thoughts with AI are are this, If I stop using AI, the environmental damage is going to continue anyway, because I can't guarantee everybody's going to stop. If I keep using it, at least I benefit and at least my contribution is insignificant. If everybody keeps using it, an I opt out, I'm worse off. So the rational individual choice is to keep using it, is to keep using AI. Right when everyone makes that choice again, whenever it makes the same rational choice, emissions rise, water demand explodes, we end up in a bad situation. Psychologically, we as humans, we know this. We're terrible at protecting resources we can't see and don't immediately pay for, so we keep going until we reach destruction. The environmental impact is one thing, but this isn't an environmental podcast, This is a psychology podcast, and the psychological impact of that impact is the one I want to dissect more. I think something we're not talking about is how the threat of AI is making so much of us, so many of us I which should say, really pessimistic about the future and really depressed. We're in this like weird reality where we all feel so hopeless. We know people are going to lose their jobs, it's unavoidable.

00:39:03
Speaker 2: We just know that.

00:39:05
Speaker 1: We know that climate change is going to get worse. We know that we're going to stop being able to tell what's real or fake online maybe even in person, and that is hard on our collective mood and collective mental well being. We're seeing these fears reflected in an increasing sense of fear, listlessness, hopelessness, anxiety, especially in young people. You know, we are not as positive about the future as the people and the generations before us. We're one of the first generations who doesn't say and who won't say that they have things to look forward to. There was a twenty twenty four study published in Frontiers and Psychiatry that already has showed really high and increasing prevalence of existential anxiety around AI, particularly a sense of emptiness, anxiety about meaninglessness, guilt over AI related catastrophes, and unethical actions of AI. Is anyone doing anything about this? Anyone doing any thinking about this? Do we have specific institutions set up to cope with this specific arm of fear and a specific type or new form of psychology to manage AI anxiety, particularly now that it's the dominant world technology. No, we don't, and I think humanity will increasingly start paying the price for this, a psychological price. Finally, we have to talk about its impact on creativity. AI isn't doing great things for our critical thinking skills overall. It's not doing great things, but our over reliance on it to make art or generate creative ideas is explicitly harming our ability to have novel, unique creative thoughts. One randomized study from the University of Toronto that looked at over eleven hundred participants I think, showed that the use of these large language models and generative AI systems reduces the human ability to think creatively, resulting in basically more homogeneous vanilla ideas and fewer truly innative ones. Art becomes boring, essays become boring stories become all the same. And whilst these tools and they did find this, they can increase and enhance short term performance, they reduce your ability to think independently over time. So how I see this as like people who use steroids for sport, or like performance enhancing drugs. Right, might make you better for a few days, might make you better for a few months. It is always going to have cost later on. It's always going to create over reliance. I saw this comment from somebody that said creativity is a skill, it's not an innate talent, and that talent is not developed to the same degree when you rely on AI and I completely agree. Being creative, thinking outside the box, generating ideas is difficult. That's half of why it's enjoyable because it challenges you and it gets you to it gets you to execute something complex and interesting. And so when we have this easy technology that can do it for you. They can write the end of the script, that can generate the start of the poem, that can generate an image for a friend's birthday card. Of course you're going to take it, especially in a world that values efficiency more than creativity, and it always has. However, every time we make that choice, we deny our brain's ability to tap into a specific kind of thinking called divergent thinking. This is a type of free flowing, undirected, mysterious thinking that generates new ideas and new pisptives and new answers. Our ability for individual divergent thought is one of the most human things about us, and it's made up of four things. Fluency the ability to develop a large number of ideas. Flexibility, the ability to produce ideas in many categories, originality, the ability to think about something in a way that others couldn't. And elaboration the ability to take an abstract thought and make it into something real, so basically, have an idea, turn it into a piece of art. If you take our ability for divergent thinking away due to disuse or replacing some part of that divergent thinking process with a machine, you take away a.

00:43:32
Speaker 2: Big part of your humanness.

00:43:34
Speaker 1: I read this wonderful Substack article that said, the most important thing creativity does for us is that it helps us know thyself. And you cannot know thyself if you ask Ai to generate ideas for you, because you are relying on something beyond you that does not have a soul does not have a spirit to do the sacred work of a human. It's treating creativity like something that needs to be efficient and productive and perfect and mass produced, when that is the opposite of what it's intended to be. When everyone ends up making the same stuff, the reflection of personhood and individuality that creativity is meant to channel is destroyed. Essentially, over relying on AI is outsourcing your muse. Also, let's be real here, there's an ethical concern. AI had to learn how to be created from somewhere because it's not naturally creative, because it doesn't.

00:44:26
Speaker 2: It's not real.

00:44:27
Speaker 1: I had to learn how to produce art and produce stories. Where do you think it learned that from? It was trained on the art and work of living people. Look up anthropic AI settlement. Look up this settlement case if you want evidence that this is happening. This company, and I know about this because I'm an author, right and I got an email about it. This company literally illegally pirated over seven million books to train its book writing AI model. It's now been forced to pay over one point five billion dollars in a settlement suit. Imagine using AI for a creative project, only to realize you copied some one else's work who didn't get paid assent. I don't think that's a good feeling ethically. So to summarize this is what we have come to. AI is costing our brains on a collective scale, their ability to think critically, their intelligence, their ability to connect with others, their creativity, and our sense.

00:45:21
Speaker 2: Of humanity and well being. Is that the end? Is that? It?

00:45:25
Speaker 1: Like?

00:45:26
Speaker 2: Is that just our destiny? We just have to accept it. I don't think so.

00:45:31
Speaker 1: I don't think so, otherwise I wouldn't be making this episode.

00:45:35
Speaker 2: I think we can.

00:45:36
Speaker 1: Have positive and have a positive vision for ourselves in the new AI world by practicing some discernment and continuing to use the skills we know could be overtaken by AI but still feel good for us to practice and make us feel human. So, just to quickly finish off this episode, I want to go through some ways to counter out to the impact of AI on your brain and psychology. Number One, let thinking cause you friction. Friction is what builds skill the way discomfort is what builds muscle. Don't think of friction as the cost of thinking or the side effect of thinking. That needs to be eliminated. Think of it as the reward you get for thinking. Your brain benefits from problem solving. It benefits from what we call productive confusion. Productive confusion is when something feels difficult in the moment but strengthens your capacity in the future. There is this tremendous twenty fourteen article by researchers at the University of Boulder, Colorado, called literally the Benefits of Confusion for Learning, which basically tested this theory. It tested this theory that things that are harder and contain contradictions and conflicts in multiple sources teach us more. And what they found was that that is true. Friction in learning promotes deeper understanding and intelligence. One way to practice that the thirty minute rule before you commit to AI, Before you say I'm an AI, this chat, GBT, this be if I'm gonna ask chat, commit to thirty minutes of searching listen. AI will always be there. But this is about values. Do you value convenience or do you value deep learning and intelligence? Do you want to be someone who can only do things quickly or someone who can engage deeply independently, And it's honestly up to you you have free will, but you just have to acknowledge the consequences of being AI dependent here. Secondly, question whether using AI is absolutely necessary in a specific circumstance, and question how much time it's actually going to save. I think we could all benefit from using AI less, so just be selective with what you use it for. I think small tasks like knowing the weather or replying to an email, you can probably do that yourself. In fact, it's probably a good thing if you do. A lot of AI researchers will tell you the best use of AI is not minute tasks. It's also not complex tasks. It's process tasks that reduce cognitive load but still require human reasoning, abstraction, and thought. So examples of this formatting, structuring, outlining, searching. When AI is used for micro tasks, it risks learned helplessness. Basically, I can't do anything without this. When it's used for complex decisions or large large amounts of thinking, it risks authority bias. You know, just it's said this is the case, so it must be true. And it also causes us to lose a lot of originality.

00:48:58
Speaker 2: You know, I'm not stupid.

00:48:59
Speaker 1: As much as I try and avoid AI in my own life because my values. I know a lot of us will use it. It's about using it consciously because you know, hopefully what the cost is, the water, the energy, the brain tax. And I think this brings us to our third tip, which is to limit how often you are allowed to use it per day. This is something my good friend Lizzie does. Obviously, she doesn't want to deny herself the efficiency benefits of AI, especially since she works in a I was gonna say where she works. She works, not gonna. She works in a corporate environment that is very demanding and everyone is assisted by AI. But what she does is she limits herself to three searches a day because it's a priority for her to keep her creative and original thinking skills sharp. She uses it as a tool, which is what it should be used as. Just find the formula that works for you.

00:49:59
Speaker 2: Tip number four.

00:50:00
Speaker 1: Notice when it's replacing human sources as well. Again, I'm not going to deny that AI isn't amazing for some things. What it's definitely not amazing for is giving you personal advice, relationship advice, therapy advice. When it's telling you how to answer someone's text when it's telling you how to communicate with another person. That is not how we should be using AI. That is something that I think is sacred, being able to solve problems between us, being able to communicate and be open and learn from each other. The moment that is replaced by a machine like I just think all help is lost because empathy is then empathy is then something that we are saying is so unimportant it can be offloaded. So those are the tips I think we need to kind of stay grounded in AI. Also, just let yourself be bored, let yourself be creative, make sure that you are continuing to tap into what makes you feel human. I think that will just make us feel more positive. It will make us feel more positive about ourselves and about the fact that we are alive and sentient and human. And also it will allow us to keep building skills that people who only use AI won't have. And I think that gives you an edge that's going to become even more important in a very AI centric world. Being creative, being a divergent thinker, thinking broadly, being eccentric. That is going to be the next biggest skill in an AI world.

00:51:36
Speaker 2: So if you've got those.

00:51:39
Speaker 1: I honestly I feel positive for your future. I think you're going to be just fine. Thank you, as always to our researcher Libby Cobbert for her help with this episode. Thank you if you have made it this far. If you have and you're listening on Spotify, leave a little computer emoji down below so I.

00:51:57
Speaker 2: Know you're a loyal listener.

00:51:58
Speaker 1: And thank you as always for for being here till the end. I really appreciate it. I hope you enjoyed this episode. I hope I didn't come off too strong. But also I'm down for a discussion. So if you have further thoughts, dm me email me. However you want to get in touch, you can follow us on Instagram at that Psychology Podcast. You can also follow us on substack and TikTok. And if you are listening and you want to watch the podcast, we are on Netflix, so you can tune it over there if you haven't yet, and if you're in the US and Canada. But until next time, be safe, be kind, be gentle to yourself, be AI conscious. We will talk very very soon.