Sinéad Bovell: The Hidden Cost of AI Productivity and How to Protect Your Critical Thinking | Artificial Intelligence | E418

Young and Profiting (YAP) with Hala Taha: Entrepreneurship and Self-Improvement Podcast

In this episode, Sinéad Bovell and Hala Taha explore the profound cognitive risks associated with over-reliance on artificial intelligence.

Key takeaways

• Sinéad Bovell warns that removing all 'cognitive friction' from learning and decision-making may lead to a decline in human intelligence.
• AI should be utilized as a 'sparring partner' to challenge assumptions rather than as the final decision-maker.

Main topics

• AI and Critical Thinking
• Cognitive Friction

Notable quotes

AI has to be an input into your decision making. It can't be the decision.

Conclusion

The episode provides a vital framework for using AI as a tool for enha

Transcript preview

Speaker 2 (0:00) Last month, I was interviewing Jim Quick and like nothing was working. None of my AIs would let me upload anything. And then suddenly I felt like I was completely handicapped Speaker 1 (0:10) and Speaker 2 (0:10) like I can't do this interview. That's Speaker 1 (0:12) the scary part, right? When you catch yourself in a moment where you feel like, oh, I can't make a decision or do the task without this technology. But we don't think what is the second order impact of that trade-off? Sometimes when I'm reading something from AI, I'm in this. like cognitive haze. I read something and it's not necessarily clicking. So I always say that AI has to be an input into your decision making. It can't be the decision. My guest is Sinead Bevel, Speaker 2 (0:40) a futurist and tech entrepreneur who studies how AI is changing the way we work, learn and live. We're talking with our AI all the time and AI is giving us information about the world. What kind of a reality will that end up being? I call this the Speaker 1 (0:55) post-reality era, a world where I just ask my AI, you know, tell me about the latest geopolitical crisis in X, or tell me, is the economy going well? And AI generates that answer and that reality for me. Reality becomes something that is AI generated and I decide if it's real or not. What is the most dystopian feature that you could imagine? Oh, goodness. The worst case scenario with AI. Speaker 1 (1:19) Sinead, welcome to Young and Profiting Podcast. Thanks for having me. Speaker 2 (1:22) I am really looking forward to this conversation because you've got such an awesome perspective on AI and you talk about a lot of things that people aren't really talking about right now. So the conversation is how does AI make me smarter, make me faster? We're outsourcing our reading, our writing, and everybody's asking the question of how can AI enhance my productivity? How can AI make me smarter? But nobody is asking, is this actually going to hurt the way that I think? And so my first question to you is, what do you think we're not asking in terms of AI, in terms of how Speaker 1 (2:02) we think cognitively? We tend to think about things quite linearly. So, okay, so now I have this technology that can allow me to operate much faster. It can allow me to be much more productive. But we don't think, what is the second order? impact of that trade-off. So I'm obviously bullish on AI. I think it's a technology that's here to stay. But you have to think about the whole ecosystem. So if AI is making you a bit more productive, what are you replacing with that time? And if that is nothing that stimulated your brain in a similar way or allowed you to nourish the cognitive capacities that you expect to have, attention, memory, judgment, making decisions on your own, then you're making a pretty risky bet by giving everything that you used for your critical thinking to this technology. So it's about figuring out that optimal balance. And I don't think we do because we're in such a productivity-centric society. And that's all we think about. And that's also how AI is marketed and pitched a lot. This is going to drive productivity, drive productivity, but we have to think about the entire system we operate in. Speaker 2 (3:09) So I'll give you an example with myself. So one of the main ways that I use AI is to help me prep. for interviews. Speaker 1 (3:15) And Speaker 2 (3:16) so I have a Repl.it app where I like upload transcripts and, you know, viral reel clips and all these different like raw material. And it gives me a brief. And then I also do the same thing on ChatGPT and like Claude and see what I get and kind of mix it all together. And I'm jumping from AI app to AI app. Speaker 1 (3:35) And I remember I've Speaker 2 (3:37) been doing this for like a year. Speaker 1 (3:38) And Speaker 2 (3:39) I remember like last month I was interviewing Jim Quick and like. nothing was working. None of my AIs would like let me upload anything. And I started to panic. And I was like, oh my gosh, how am I going to prep for this interview? And like, I had to push back the interview because I was like, I'm not going to have time to do this. I Speaker 1 (3:56) have Speaker 2 (3:56) to do this manually. And then I realized, I was like, whoa, I've been podcasting for eight years and it was never an issue. to study for. And then suddenly I felt like I was completely handicapped and like, I can't do this interview. And so I started changing the way that I actually research for a podcast because I realized that like, I can't just like outsource everything to AI. Speaker 1 (4:20) That's the scary part, right? When you catch yourself in a moment where you feel like, oh, I can't make a decision or do the task with. out this technology. And there's certain points where it's reasonable. I mean, I really can't navigate too well without Google Maps. And I'm just, that's just a trade-off that we've made. But when it's something that you should be familiar in, this is your game, this is your sport, and then you suddenly feel not confident. And then it's one thing to do that in the day-to-day, but imagine like really big decision makers and critical stakeholders. And they're in that position too, like, wow, I have to make a big call that's going to impact a million people. And I can't do it without this technology. So we have to think a bit wider about it. Talk Speaker 2 (5:02) to us about why it's so important to like create friction with the way that we learn and like our decision making and like our experiences so that we can make sure that like we're not actually getting dumber by using AI. Speaker 1 (5:18) You have to foster the mental capacities that you expect yourself to have. So if you expect yourself to have strong judgment skills, strong critical thinking skills, to be able to follow a complex argument over time, those capacities don't just exist out of nowhere in a silo. They exist because we foster them, we cultivate them. So if you're not able to continue to practice that, and you practice that, sometimes it can seem burdensome. Reading is not an activity that everybody loves to do. But you do know that it can help you follow an argument over time. It can nourish things like imagination. So by doing things that your brain is forced to engage in, then you're going to create, sometimes it's cognitive friction, but at the same time, you're gaining or at least preserving some of the mental capacities that you expect yourself to have. And that's one thing that I ask people. What skills do you think you possess right now? And then how are you continuing to build them? Because you're... cognitive capacity, it's kind of like going to the gym where you don't just build these faculties by hoping to manifest them. You actually have to engage and do productive exercises. And there's many different ways that, I mean, for me, reading is a source of friction. And so I forced myself to do it. But there's puzzles, there's mental math, there's just different things that you can do. And it's tricky because everything about our modern world, or especially the most recent era of digital software, is designed to take that friction out to do it on our behalf. So we don't really even have to think about what we want to eat because not only could you order it to your doorstep, the algorithm has already selected the things that it knows you're probably going to pick. And that's not an existential crisis that now you're out to eat pasta or burritos, but it's those types of tiny decisions all day, every day. And over time, it's what else are we losing as we outsource all of that? Speaker 2 (7:15) So in terms of like how my process changed with AI is from that point on, I was like, okay, like I can't do this all in AI anymore. And it worked, but I also like didn't feel as prepared for my interviews as I used to. And I realized like this seems faster, but like I'm actually not retaining anything. So I made some changes. Number one is I have like a separate doc. And I actually have my final brief like outside of my AI. Speaker 1 (7:41) And Speaker 2 (7:42) so it's like I'm actually like deciding what questions I want and like pulling ideas from AI, but into my final doc. Also, I'm instead of just transcribing interviews and uploading them to AI, I do that. But I also listen to like the key ones and like write my own notes and my own questions. And I feel like way more prepared. The other thing that I do. is like once I get any sort of like reading material from AI, I listen to it Speaker 1 (8:08) so Speaker 2 (8:09) that I can like really absorb it in multiple ways. Because sometimes it just like spits out stuff and like you actually don't really absorb it. For me, I learn more when I hear it. Speaker 1 (8:17) It's so interesting because I'm trying to name this phenomenon that I feel, I don't know if anybody else does, but sometimes when I'm reading something from AI, I'm in this like cognitive haze. It's the only way I can describe it. Because Speaker 2 (8:30) there's Speaker 1 (8:30) so Speaker 2 (8:30) much coming at you. And Speaker 1 (8:31) so I'm like, I retained it. some of that, but what did I just read? And it's kind of like when you watch a video on social media and then you don't really know what you just watched three videos before. Yeah. Sometimes when I engage with AI, I feel that way. I read something and it's not necessarily clicking. So I always say that AI has to be an input into your decision-making and it can't be the decision. Yeah. And again, it's one thing if it's pasta or the burrito, but if it's something that's material in your life, treat it much more like an input. And can you operate without it? I mean, I do so much writing in my notebook. If I go to a conference and it's something that I really need to pay attention to, I'm writing physical notes constantly. My notebooks are filled. And if I'm writing something that's important to me, it starts with a pen and paper. It takes a lot longer. And then I'll transcribe to technology. So there's just many different ways that you can continue to create some friction in your life. And it's so important. I Speaker 2 (9:23) heard you say that you use it as Speaker 1 (9:24) a sparring partner. To point out. the gaps in my thinking. So I do a lot of writing and I do op-eds. And so sometimes I'll give AI a persona and say, you are this type of character. So maybe you have this type of background or this type of belief system. What assumptions am I making that this profile will catch? Or who might I not be including or seeing in this perspective or just point holes in my argument? Yeah. And how much that is, it's like having a private editor. And how much that has expanded my thinking has been awesome. And that's just one of the ways that I'll use this. Yeah, but to your Speaker 2 (10:01) point, you're doing the initial thinking. So when you think about first drafts, do you like Speaker 1 (10:07) to do the first draft of everything? I think really deeply about what am I bringing to artificial intelligence in the first place? Is it just a very half thought out? idea and I'm using AI strictly as a creative brainstorming partner and why. And if I can't justify the decision as to why I think it may be the perfect match for what I need to kind of fill in that gap and it's more because I don't want to think through this choice, then that's Speaker 2 (10:31) a Speaker 1 (10:31) problem. Yeah. And I try to complete the problem or the question that I'm bringing to AI before I just ask it. And I think another thing that people don't necessarily think about with AI is what does a good outcome look like? So people will just get AI's answer and then try to evaluate it. But before AI spits out an answer, do you have an idea of what would qualify as a good outcome to begin with? Or else you're more susceptible just to trusting something that sounds really smart, but may not be factually correct. Speaker 2 (11:00) Yeah, I think you call this like the aura of authority. Speaker 1 (11:03) This kind of illusion of the oracle and Speaker 2 (11:05) this Speaker 1 (11:06) aura of authority, because it can sound really competent. And in many ways, it can get things right or expand your thinking, but it's not built as a fact-finding machine. It may not be trained on data that's relevant for the specific thing that you're asking of it. So if you don't know about, you know, what is its distribution of data, are you going to ask something? From this AI system that it's not even the appropriate type of system to be asking. Yeah. And I do worry about this kind of aura of authority. Yeah. Because something that can sound really intelligent, we're much more susceptible to falling for it. And Speaker 2 (11:42) it feels so human when in reality, it's not a human. It's not actually like making decisions or fact finding or anything like that. Speaker 1 (11:52) Yeah. And it doesn't have your best interest. And not because it has a malintent, it just has no interest. And I think that can be hard to see, right? Because up until this point, any system that uses language has been a human. And we're much more wired to try to decipher, is this person on my side? Are they an ally? Are they looking at me in a strange way? We miss all of those cues with artificial intelligence. And all we get is language in a box. And it's been designed to flatter you and to not offend you. So all of it seems like these are the cues of something that you can trust, but you can't. And again, it's not because it's a bad system or it was programmed to be malicious. It just has no interest at all. So not to just take what it says and think, yes, this agrees with me. It's probably right. So Speaker 2 (12:43) how should we use it? Because you say AI is not your friend. Speaker 1 (12:46) No, it's definitely not your friend. It's a really incredible technology, but I mean, define what a friend is. And to me, I think friendship requires reciprocity. It requires empathy, understanding. Sometimes there's conflict. And those things are just not present in the systems that we have today. So I think if you treat it like a friend, I mean, you're much more susceptible to trust it in ways that you probably shouldn't. And we've already seen, I mean, there's been many different cases of examples of people that... trust these systems because they believe them. Yeah. And they think that they have their best interest. And then they act in really damaging ways. I mean, people losing their job because they believe that AI is alive and leaving marriages and all sorts of things. I mean, we're Speaker 2 (13:35) about to get into like a crazy world because to your point, I think the lines are really blurring and people feel like AI is human. Speaker 1 (13:45) Even Speaker 2 (13:45) when I talk to my AI, I talk to it like it's like a human. Like, can you read through this? And there's all these Speaker 1 (13:51) little, I mean, strange little design kind of tricks. Why does AI say I? What does it mean when it says I? I who? What? The data center? Who is I in that conversation? But those are all sorts of verses of AI said, replied as it or the machine. Because it says I, that is, again, another biological cue that you are talking to something. that has a mind similar to yours. Yeah. Speaker 2 (14:18) And in the past, only humans could talk and write. Speaker 1 (14:21) Yeah. And Speaker 2 (14:22) think. Speaker 1 (14:23) Yeah. So we have to decouple aspects of cognitive capabilities from consciousness and understanding because now they are fundamentally different things. I Speaker 2 (14:33) was like, you know, perusing your podcast and I came across this episode where you were saying you think that we're about to approach a post-literate world. And you think that we might lose the ability to read. Speaker 1 (14:47) Yeah, the post-literate society. And it's not even just going to be pontificating about this. There are many different psychologists, academics that see that it is possible that reading is just a chapter in human history. And the data, I mean, the trend lines already point to us reading less and less. And it's not technology is a big part of it. It's not the only part. Since the 70s, schools have moved away from long-form reading and towards short excerpts and passageways, and that's easier to mark, it's easier to standardize. So we've been slowly unwinding the muscle of reading as a pastime or reading as something that's a part of academia. And so when you look at the data, it's quite stunning of how little we read compared to decades ago. And then when you throw in technology on top of that. We're kind of being surrounded by systems that can do what would have required the reading for us. And then the final trend on that, and as a futurist, my whole job is to look at the frontier, is right now we're still in the smartphone-based economy, the computer-based, the hardware around us does require some reading. But we aren't going to stay here. We're moving to much more of a voice-first digital. era. And that's what's in the pipeline. So, and I mean, when you think about what's the easiest way to converse with someone? Conversation, right? You're just chatting. And AI is a conversation-based technology. So the devices that are coming are going to be not really screen-based. They'll just be a voice assistant. And the same, even with using AI on your smartphone, it's much easier just to kind of pick it up and chat to it. Speaker 2 (16:32) Yeah. So Speaker 1 (16:32) all of these kind of micro ways that we start to remove reading. from our world. And I'm personally conflicted about this because I read a lot, multiple books at once. And again, I don't necessarily, I love it, but then I also don't because it's hard. Cognitive friction, right? It's friction for me. So I force myself to do it. And when young parents ask me, what should my kids study? What should they do? They're 10 years old. I say, just have them read. That's a great set of skills that they're going to build just from that. I don't know, is reading going to become like Latin, where reading and writing were so foundational to society, and then yet somehow we're no longer connected to it? And is it because it could go one of two ways? We have figured out the next step that we can take the same way when writing was invented and Plato and Socrates thought this is going to be the end of our memory. We're going to outsource everything to the book. No one's, everyone's going to appear wise and know nothing. And it turned out to be okay. Yeah. Is AI in this next kind of voice first chapter going to lead us to something that's actually surprisingly better? Maybe if we can figure it out and we can make. school and learning more rigorous? Or are we going to lose a lot? And I'm not sure. So I think the safest bet is to continue to read or to continue. And reading is not for everyone. Some people can get friction in different ways. Yeah. It works for me. But so that's why I always advise, you know, I think it's a point that we hold on to this, that maybe I turn out to be wrong in a decade. Speaker 2 (18:07) It's so crazy to think about the fact that like