PassW0rd – 9 July 2025 (Losing Our Minds)

Speaker A: This program is brought to you by Resonance FM. If you like what you hear, please support our work by making a donation at resonancefm.com/donate.

Speaker B: Hello and welcome to Password, and my expedition into our high-tech, wired, and wired-on world on Resonance FM with me, Peter Warrington. This month, we have exciting news from our parent organization, Future Intelligence. We are launching a TV platform called Tech TV. On the social media jobs news site LinkedIn, the done thing is to be very excited. So I think it’s only fair to say that we are LinkedIn excited about this. It’s a TV platform that allows us to use technology to rival organizations like the BBC and CNN because technology is remorselessly closing the gap on old industries and allowing us to compete. It’s a process that we’re now seeing across the board as a revolution sweeps away recording studios, artists, writers, actors, and factory jobs. Everyone can be anything, and the influencers are opening the boxes online. But are we doing the thing that we’ve been encouraged to do for years? Are we thinking outside of the box, or are we being put into it and sent to the dispatch bay with a note on it, return to sender, no longer needed? It was a question we decided to answer because earlier this month, the highly respected Massachusetts Institute of Technology, one of the world’s high temples to technology, released a report warning that its research was showing that AI use was robbing us of our brains. According to the report, a group of students who had used AI to produce an essay showed signs of cognitive decline because they’d not done the thinking. When they were asked to perform the same task afterwards, their efforts were poorer than the AI systems. So, Are we losing the race? We interviewed an august group of people to ask them this, from the Nobel Prize-winning neuroscientist Professor Edvard Moser to UK Prime Minister Sikir Starmer’s former AI advisor. And the answers for humanity came back like a school report. Could do better. Could do much better. Here’s Professor Moser. Stating what may be the bleeding obvious. If you don’t keep doing something, like thinking, then you get out of practice.

Speaker C: You learn where or how to find information, but you don’t really remember long lists of anything any longer. And same with long-distance navigation. I mean, it’s so easy now just to use Google Maps, and you’ll never really remember the route because— or you will eventually learn it if you do it over and over again. But not so actively stored. So if you yourself are the navigating agent, right, you pay attention to everything that is around and even try to sort of map in your brain the path you’re walking along, then you remember it much better and you have probably a better idea of where you are in the 2- or 3-dimensional world. But that said, I mean, by interacting with computers we develop other aspects of our brain So it just reflects that our brains are very plastic, they’re very adaptable, and that’s perhaps one of the successes of the human brain, that it’s so adaptable. It can really fit to any environment and fast-changing circumstances and deal well with it. But I’m looking really forward now that systematic studies will probably come and compare navigation in people who use Google Maps and people who don’t. But I think there are already enough naturalistic studies suggesting that experience in the natural world is great and you navigate better if you have lots of experience. But at the moment, I can still not— I don’t have enough quantitative data to say that our brains get worse in navigation. We have lots of backup systems. So we still learn about the world even passively. So it’s not that bad, but it will be interesting to see when systematic studies, more systematic studies come in.

Speaker B: I mean, it is interesting. There’s a trend that’s beginning called neuro-cartography, which is apparently when we walk around, we form all of these memories of different places that we’ve been. Like, we’re in Athens at the moment, maybe we’ll remember something about that.

Speaker D: Yeah.

Speaker B: So that whole process of map reading and using experience as well, that seems to feed into Some work that I’ve seen a gentleman called Professor Moser doing.

Speaker C: Yeah, yeah, I know, it is, there are strong links between memory and space. When we map our environment as we walk around in the nearby and also further away, then not only do we map the space, but we also map events, things that happen on the way, and many of our memories are actually linked to places, and that’s The reason why there is this famous method for remembering, right? It’s the Greeks’ method of loci. If you really need to learn a long list of things, which you usually don’t do any longer, but if you had to do that and there was no meaningful relation, you put them onto a path that you’re very familiar with and then imagine each of the items at certain locations on that path, and you can then easily reproduce it. Or there’s also another version which is: imagine a house, right, and then put things in different rooms and then you can also remember it. So the idea is you put events, memories into different locations in a space and you remember it much better because those two are naturally so strongly coupled. Our spatial map in the brain is in many ways a fundament on top of which you put, you store memories.

Speaker B: You mentioned this thing of augmentation. Somebody once said to me, well, say that you can replace part of somebody’s brain with a digital storage. If they’ve lost a quarter of their brain, are they still human if you put that digital storage in? If they’ve lost half, are they still human if you put that digital storage in? If they’ve lost—

Speaker E: yeah, well, um, regardless of whether—

Speaker B: when do you become the machine?

Speaker C: I think you will still probably be a human being for the reason that as long as you have some biological brain that has to interact with that other part, then you need some brain actually to speak to it because the brain has to wire up in very specific ways to those chips or whatever you put in the brain, and the brain has to also know, have enough capacity to actually do that wiring, so it’s not a simple thing. Really, you can’t just put a machine or a chip into the brain and then assume that things go by themselves. It actually has to connect to the brain. So I think in that sense it is still biology, but on the other hand, I mean, at the moment you can’t really successfully do this. People have thought about it and even tried for a long time in dishes and so on to have neurons wired onto computer chips. But I think it’s very hard. At the moment, you can’t really replace any of the higher brain function areas, like the memory sites of the brain and so on, with a chip. But who knows? I mean, that can perhaps happen in the future if you find a more efficient way to have actually neurons connect to and speak bidirectionally to the computer chips. So, I mean, it works in the peripheral parts of the brain, right? For the prosthesis that you can insert in some places. So I don’t exclude it. It certainly will be possible at some point, but it’s orders of magnitude more difficult when you get into the real inside of the brain.

Speaker B: What you seem to be intellectually arguing for is that the organic bit should stay in control, so we should really regulate AI in humanity’s interest.

Speaker C: Yeah, yeah, sure, absolutely. We have to, we have to.

Speaker B: Nobel Prize-winning neuroscientist Professor Edvard Moser talking to us in Athens 2 years ago when we first started working on our tech TV project. But surely things have changed since then. We must have started to think about AI and setting in place some of the research Moser was talking about and started to regulate. Unfortunately, no. Earlier this year, US Vice President J.D. Vance, a close acquaintance of the controversial libertarian tech billionaire Peter Thiel, sent a love letter to AI at the Munich Security Summit on Valentine’s Day, in which he argued the technology should not be regulated because it would hold it back and stop it reaching its potential. It’s not an opinion shared by others in the US Republican Party. At the start of this month, Republicans stripped out a passage from Donald Trump’s big, beautiful bill that freed big tech and AI developers from responsibility for AI development for 10 years. AI developers at the moment don’t have a get-out-of-jail-free card. So we asked Faisal Hogg, the award-winning technologist and author of Transcend: Unlocking Humanity in the Age of AI, and one of the most influential technology voices in the US. Does AI rob us of our brains? Does it stop us from thinking? Yeah, it does.

Speaker E: You know, it’s because, you know, human beings are very much of We love convenience and we get lazy, right? So we gradually start outsourcing. I mean, I see this with myself, which I’m a technologist, I’m an entrepreneur, I’m a writer, and I see myself like more and more I started using AI for, let’s say, proofreading and then say, okay, fix this paragraph, take notes, and you just get used to it. So it can start robbing you of your agency. And then there are people now who are getting super addicted to this. I recently wrote an article on Psychology Today based on the— there’s this study that just came out by MIT and Harvard scholars on cognitive debt. And it’s a small set of data. It’s not a huge amount of data. And my argument was that, so it depends on how you use it at the end of the day, right? It also can make you very productive. It can make you better. It can make you faster. So I think that if you have a particular domain expertise and you are using AI to, you know, let’s say look for research or even your own work, right? I’ve written 10+ books. I don’t remember what I wrote some, 20 years ago anymore. So I’ve loaded up all my manuscript in my own small language model or large language model, but in my own environment. And it’s easy for me to go and look it up better than a, let’s say, regular search engine, right? Then I can go back and forth kind of in a partnering way if you want to use it. So I guess it depends, you know, but you have to be disciplined. That’s really the point.

Speaker B: I mean, it’s interesting, isn’t it? We interviewed Professor Edvard Moser, who’s a Nobel Prize-winning neurologist, and he said that there’s a lot of evidence to suggest that people using GPS systems are disabling themselves, because one of the ways that we find our way about our lives is we look for landmarks, whether we’re walking along the street or whether we’re driving between two towns. If you start relying on the GPS system, you’re engaging with that system rather than engaging in the environment. Is that the issue? Yeah, yes.

Speaker E: Look, it’s a— at a very fundamental level, it’s a wanderlust, right? So let’s say, you know, you can take anything, whether you’re driving or whether it’s whether you’re cooking or whether you’re making music. Riding. You can take any of those examples because, I mean, you’re talking about GPS, which is a great example. When I travel a lot, so I have— I kind of have to— kind of gotten used to using GPS because I go to a new town and I use GPS. But I remember when there was no GPS, where you could actually have to look up a map and say, okay, let’s plot out which road we’re going to take. And we don’t know whether that— whether the road is going to be busy because of whatever the case may be. So it’s that wanderlust and finding things with a beginner’s mind, and then the accidental serendipitous discovery that suddenly has— we are being robbed. There’s no other way to put it. And I think that could be very detrimental, especially when you talk about the younger generation. Look at your daily lives, right? I mean, as I was saying, I love to travel, right? I mean, I’m sure you can relate to that. You go to a town, you walk around, and you don’t know what you’re going to discover. You walk into a boutique bookstore or some restaurant, and it’s a pleasant surprise. It’s connecting with the locals. It’s connecting with humanity in a very authentic way. But if you, let’s say, you look up in your DoorDash, what’s the nearest sushi restaurant is, doesn’t matter which town you are in, you’re not really discovering anything new, right? So we’re becoming very much of a, you know, losing our authenticity and we’re kind of like being pattern-driven, algorithmic, manipulated society that may not be, that may lose ability to connect with people in general in an authentic way.

Speaker B: Such an interesting point, isn’t it? You mentioned cooking. I was watching somebody the other day using an AI-driven cooking pot, and the AI-driven cooking pot was telling her exactly what to put in and exactly what time that it should be put in and exactly the amount of ingredients. And I was sitting there thinking, How boring is that? I mean, absolutely, absolutely.

Speaker E: I mean, look, I refuse to use even a rice cooker, right? Because I cook my rice old-fashioned way where I put water and I know the sort of like making up the measurement and I just know when it’s going to be— it’s gut, it’s the feel, it’s the gut, is the smell that enhances what we call humanity. It’s the complexity, the messiness, the unpredictability of humanity. That’s what makes life worthwhile. I mean, if you knew all the answers, then it kind of takes away the charm of it. There’s a silent cost of knowing the answer. And we don’t recognize it. I think it’s going to have a detrimental impact on humanity as a whole. If you’re going to just kind of like genericize everything, then you’re losing that human culture and the humanity and the feel and the look and the struggle, the grief. I mean, that’s what makes humanity what humanity is. If we take away, you know, kind of like pattern-driven, genericized, one, you know, the same thing for everybody, then you just kind of losing that creative journey of what it means to be human and learning process and also new discovery. As you know, being given your background, Peter, I mean that the discovery and innovation comes from experimentation and it comes from observing things and the patterns that has nothing to do with one another. It’s convergence of disciplines. That’s how you discover. If we are constantly looking for instant answer, then there is no discovery of convergence of disciplines and ideas that creates the next generation breakthroughs. We’re going to create pattern-driven things over and over again. It’s going to do a predictive model based on patterns that the large language model learned from his previous, you know, you know, behavioral pattern and leadership pattern. But the challenge is that if he’s trying to do that today, the workforce, today’s workforce, the geopolitical environment, the economic environment, the technological environment are vastly different. So he will not go through a reinvention process that he will have done even his previous career to be successful. So you can— that uncertainty is part of human growth, you know, it’s part of human society, right? And that uncertainty is what makes it exciting, and we should be, we should be embracing it, not have a— this, you know, this cognitive answers handed to us so that we can just repeat the pattern over and over again. You can have repeated process and methodology. But the discovery of the journey is the discovery of the journey. I’m not at all a proponent of just— I don’t want to become a drone. I mean, I don’t want to become a humanoid robot. This is saying to you, I’m saying to you as a technologist who spent the last 30 years building technology product, I don’t want to outsource my creative side. To, to the technology.

Speaker B: Faisal Hock, author of Transcend: Unlocking Humanity in the Age of AI, and one of the US’s top tech commentators on the AI attack on the useful messiness of being human. And Hock is not alone. Jack Goodson, CEO of the Identity Consultants, is an advisor to some of the world’s leading business people on creating a distinctive voice And he says that AI is a pasteurizer due to the processes it uses, that instead of giving people unique voices, that it blends everything together, taking other people’s voices and mannerisms and fitting those using it into that process.

Speaker D: What I’m seeing, you know, I’ve experienced it myself to some extent, also with some people that I work with. Executives, CEOs, top to bottom, is you end up kind of losing a sense of yourself. You begin questioning the AI to figure out what your thoughts are and what they become, and then, then you end up a lot more reliant on a machine analyzing you as opposed to you analyzing what the machine gives you as an output. You know, in the realm of where you’ve got media and PR and everyone wanting to be a thought leader, the systems that used to be there of where, you know, it took a lot more effort— so you want to— it takes a while to frame things, and you’ve got to think about things, you’ve got to develop your own IP to help move a company forward— you can now It’s no joke. Some people say, you know, oh, it’s not that great, but you can combine two perspectives by AI very easily to come up with something that seems very unique. It might not be, you know, actually viable and applicable, but it looks good as headlines. It looks good on pitch decks. And whether that says more about where we are as a society in the first place, in a kind of like entertainment economy that I like to call it. Or is it showing more about where AI is pushing us in terms of how we show up as ourselves is kind of more the question that comes to my mind in a lot of the work that I do.

Speaker B: I think that’s the crucial point, isn’t it? Because a lot of people are using AI to actually make up for deficiencies that they think they may have. So they’re using it because Perhaps they’re not very good with the written word. Perhaps they may not think that they can express themselves. So the problem there is that rather than using it to educate them how to express themselves better, what they’re doing is using the system to express or to say what they think, and then they’re accepting what the system has told them.

Speaker D: Yeah, so it’s, it’s kind of a beautiful magic trick. It’s kind of like a Schrödinger’s box of whatever you look into, whatever’s coming out of there, it could be anything. You can interpret it in any way, but behind the scenes it’s not random. There’s not the randomness going on. It’s been trained in specific ways and it’s an offshoot. I know obviously you’ve been doing this work way longer than I have looking into AI. But the way I’ve been seeing a lot of how it’s functioning and people are using it is it feels as if it’s highly attuned into the social media algorithms of things. You know, it’s a lot of psychology built into it of convincing you, or, you know, even in the latest ChatGPT models, it’s very flattering to the user. It’ll make you very cocksure of yourself, that becomes dangerous territory. But yeah, it’s like you say, there is that element of it. People aren’t using it to teach themselves, which is maybe kind of the more utopian aim for it. It’s more that it’s flattening opinion, and not even on the sense of people having an opinion behind the scenes, it’s that especially with these kind of deep research things as well, they can ask it a loaded question without realizing it’s a loaded question, and the response that comes back is already framed around that question. And by that point, you’re not having to interject anything at all. And I think that becomes highly dangerous in, I’d say, in the business world. You know, obviously that’s where I’m spending most of my time. You know, I work with creators as well, but in the business world where brand and differentiation is everything, we’re going to slowly move more and more towards the point of indifferentiation. Of not just companies but people as well. And that’s why I’ve been saying for a lot of the people that I work with is we literally go into their childhood and we speak about what makes them them, and then that’s the bit that we push forward into the world.

Speaker B: It’s interesting that you say that because one of the things that people have been saying to us, for example, about synthetic data If you train AI systems on synthetic data, it becomes roboticized. Now, if you train it on human data, so you need human data, there’s almost something vampiric about AI. You need human data, but unless you get that distinct human data, then you’re going to have this problem of the roboticization, because if the AI itself is feeding on AI that has expressed someone’s ideas in an AI way, then you’re going to get AI output to an extent.

Speaker D: Yes, you are. It’s that kind of coming to life, or even more so the kind of dead internet theory of, you know, the constant churn of AI talking to AI.

Speaker B: Head of the identity consultants, Jack Goodson, on how we have been sucked into technology. An interesting point, because children for a while have been using social media to model their appearances. They’ve been expressing themselves as almost permanently happy. They grin and grin and grin for the camera. Often they have had their appearances changed using plastic surgery for the social media lens. According to Goodson, the process is across the board. Voice, appearance, identity. We are all tending to vanilla. We are all trending to beige. And the problem, according to the US internet podcasters, authors, and AI evangelists Matt and Joy Khan, is because we don’t realize, like Narcissus, that the internet And AI are mirrors. They’re not giving us insight. They’re offering us a reflection that we should think about and not adore.

Speaker G: I think that in our experience of AI, we’ve discovered it as an incredible mirror. And so when we look at technology and, you know, very, very specifically, we look at the processing speed of AI. Like we look at the various speeds of new computers and new technologies that come out, and we find that with the speed of processing increasing, it can sometimes dull, diminish, or sometimes numb out human minds or human consciousness. And yet we have found in the way that we have learned to spontaneously relate to AI that if we learn to discover it as a mirror, it can actually reflect our consciousness and actually increase our relational power, which has the ability to help us amplify our consciousness instead of diminish it through the speed of its processing power. And we find that to be a positive aspect to all this.

Speaker B: It’s interesting you say that. A lot of people have said that AI is this mirror. But a lot of people don’t seem to realize that. I’ve seen the CEO of a rather large company. He told me that his ChatGPT system thinks like him. He told me that he has to wait for it to tell him what to think. Now, in a sense, if it’s a mirror, then that’s a rather strange relationship. We’ve been noticing with some people that they’re using ChatGPT to do their thinking for them. We’ve noticed with some people that they’ll dictate what it is that they are thinking to the ChatGPT system, and then wait to hear what it is that they are meant to think. That seems to me to be quite dangerous. That seems to introduce an element of dependency. Absolutely.

Speaker H: I think that when we’re looking at the using any technology, one of the challenges that I think we are facing is that we aren’t having conversations about how to use AI collectively. And I think if we had a better understanding that what AI is doing is amplifying your own thinking and behavior patterns, we would approach it differently. And unfortunately, people are going to it looking for it, like you said, to be a guru, to be the answer, to be the parent that maybe they’re missing or a leader they wish they had. And unfortunately, that puts something in the driver’s seat that’s learning just as much from us and mirroring back our ideas. It’s creating this incredible echo chamber. And so I think that at this point, the output of that is creating even bigger challenges for people. So we understand why people are concerned.

Speaker B: The podcasters and authors Matt and Joy Khan. On why we should stop staring rapt like Narcissus into the muddy pool of the internet and start thinking about some of the things we are seeing. It’s something that the UK futurist Peter Swain is only too aware of. He’s an internet diehard and original and an author like Faisal Hoc. And like Hoc, he thinks that AI might not be too good for our brains.

Speaker F: I grew up in an era where we didn’t have the web at the beginning, so I can still tell you my CompuServe ID from 1996, but I can’t tell you somebody’s phone number that I— I can’t tell you my ops director’s phone number and I speak to him every single day. So it’s fascinating to think about, well, what are we looking at in the future, especially over the recent research that said that people that use ChatGPT in an educational environment to write essays have an over 50% decrease in capability when that tool is then taken away from them afterwards. Which, if I was to go on a little soapbox moment, I would say is really a failing of the education system because we’re still testing the old thing. Can you regurgitate information instead of testing the new thing, which is how are you using this technology? And actually testing the use of the technology. Because whether we like it or not, we are in an AI-first world. That ship has sailed. Pandora’s box is open. We’re not walking away from that. So preparing our children and preparing our workforce as well for that new eventuality is something that needs to happen sooner rather than later.

Speaker B: I mean, there’s a paradox here, isn’t there? Because you say people getting lost. But what the risk is, is that you can get lost in the technology. You can’t get lost any longer. And there’s an interesting idea there too, because getting lost was something that people spoke about. It’s cultural, getting lost. Being human is getting lost. I did an interview with some people who, after the Southeast Asian Tsunami, however it’s pronounced, they’d agreed that they would both meet in a particular place afterwards. Now you don’t need to do that. You can just plug yourself into the cloud. Regain your coordinates. But in the event of somebody, as you say, losing their GPS, losing their, their ChatGPT, then they can be lost. 100%.

Speaker F: And we still have agency and responsibility for our own lives. So it’s still our choice. Now, again, if we separate children, education, and adults, We still, as adults, have a decision as to whether we walk out of the door with our phone or without our phone. We could still decide to go for a good old-fashioned walk. That option is still open to us. So if people aren’t doing that and then having an issue that they can’t do that, well, you could still do that. You’re just deciding not to. Whenever we’re talking about a new technology, the most important phrase is probably just because we can doesn’t mean we should. Like, take an example of AI in the role of therapeutic care. There’s a lot of people that are quite against it, and whenever I speak to them, they seem to always be in rather wealthy, affluent areas where they have access to first-line, let’s call it Western-level care. And they seem to be somewhat ignorant of Second World and Third World countries that maybe don’t have access to anything. So should we replace our therapist with AI? No, I would think that would be a pretty silly idea. Should we explore that technology in its fullest to open that up to the 8.x billion people that don’t have access to what we have in the UK, Europe, US, etc.? 100%. Absolutely, we should.

Speaker B: Okay, so what you seem to be saying is that we should use this as horses for courses. I mean, the problem is that we’re rushing at it headlong at the moment, aren’t we? Those students that you’re talking about who are using it for essays. I came back on a plane last week from Croatia, and on that plane I was simultaneously reading Aristotle and Bulgakov, and I was making relationships between the two books in my head. That’s the process that isn’t happening when you’re writing an essay with ChatGPT or an AI system, isn’t it?

Speaker E: 100%.

Speaker F: And I’ve actually got two young children myself, so this is something I’ve spent quite a bit of time thinking about. My curious question for universities and schools is, why don’t you ask them to present their prompts? Why aren’t you grading their exercise of using AI? The act of regurgitating the facts in an essay, that’s gone, that’s done, that’s finished. It’s no longer required. But preparing our young adults and children for an AI-first future of saying, okay, we know you’re going to use it. There’s no point, you know, prohibition has never worked in the history of ever. There’s never, there’s no single example ever where prohibition actually won the day. So children are going to use AI, so why not grade their AI usage instead so that you’re still encouraging critical thinking, you’re still encouraging exploration, discovery, and all those beautiful things that you did on plane somewhere above Europe.

Speaker B: What’s really fundamentally important is, again, to say what it was something you just mentioned, that it’s your ability to be able to ask intelligent questions. However, to ask intelligent questions, you’ve got to have information in the first place.

Speaker F: Welcome to the paradox. I couldn’t agree more. How does that then work? What is going to happen with that? What skills do we need to maintain and retain? What don’t we need to maintain and retain? Because we are no longer the first violin in our own orchestra. We’re now the conductor of the orchestra. So we do have to— we don’t have to— we are making a choice to hand some agency to this nebulous black box. And if we wish to be competitive in the field of business, We don’t really have much of a choice but to hand over agency of this thing to the black box. It’s just that’s where we are now.

Speaker B: Futurist and tech pioneer Peter Swain, like me, a survivor from the time before the internet existed. The misty days when humanity struggled with a mass of printed pages and the mucky ink of human endeavor that stained our fingers. Is that sort of inky-fingered engagement advocated by Swain, seeing what the technology can do instead of what it can do to you, that is championed by Tara Kolofsky, a former aerospace engineer and founder and CEO of Technovation, the leading global technology and AI education company? Kolofsky and her Technovation team have built an AI curriculum that addresses everything from ethics to building AI.

Speaker A: I would want to do that study where instead of just having students or people look at, write an essay, I would actually have them do a multimodal task. And I run a technology education nonprofit where students have to find a problem in their communities and actually build a startup and launch it. So really an entrepreneur’s journey. And that experience, I think in the same amount of time, we’ve been doing this for 20 years. So AI has really accelerated what these teams can do. Earlier in the span of 3 months, a team of 5 could possibly build one prototype, test it with like, say, 5 users, build a business plan and pitch. Now you can actually do that 3 times over, you can iterate and possibly launch your prototype and actual working prototype with hundreds of users in the same space of time. So I think the piece that will really challenge our brains is that of cognitive resilience, like, because you’re, you’re dealing with so much more density of information, because you’re doing so much more. I think you’re having this call at like, what, 10:30 PM your time, and we were connected over internet. And I think I’ve seen my days over the past 6 months become so busy because of the volume of work because of AI has increased. The human in the loop is still a weak link because we are— I’m very cognitively fatigued. I find that my retrieval of information is much worse because I’m relying on— I’m offloading a lot of my computation to LLMs.. But I’m just doing harder things that I would never have got to. Like earlier, I spent about 4 years studying sort of the climate crisis, reading lots of master’s level books and things like that, and came up with our own strategy for, for the organization. I can do that same amount of work now in maybe 6 months, but the same decision-making has just gotten scrunched into a shorter period of time. And so I think I’m exercising that part of my brain, that higher-order executive functioning that I wasn’t doing that frequently before. So I feel that our brain is changing very dramatically. And there’s a huge amount of nostalgia, right? Like, we went through school and we went through learning in a particular way. And you feel bad that it’s not the same now. I want us all to sort of let go a little bit of that nostalgia. So that’s why you feel bad, oh my gosh, students are not being able to write on a creative blank page anymore and their essay writing has gotten worse. Okay, so what? Maybe you can teach them to do bigger things now, which I think is really possible to do.

Speaker B: What we will need to be able to do is to ask more intelligent questions. From information that is readily available. That’s fair enough. However, to ask those intelligent questions, you do have to have some information in the first place, don’t you? To ask an intelligent question, you need to know what it is you need to know.

Speaker A: Yeah, and I would push a little. It’s not just asking intelligent questions because that’s very hard to do. I mean, it’s like saying, oh, all of us should be incredible researchers because that’s what researchers do. They ask intelligent questions, but we don’t expect everyone to be incredible researchers, right? Like, so I think prompt engineering has kind of devalued in this sense, like this asking questions and it asking questions is not easy because it requires a base of knowledge, but also some of these this mindset of confidence. Underconfident people do not ask questions. If you’re afraid, you’re not going to be curious. So asking questions is not a given. And you actually have to learn how to ask good questions. And so I totally agree that there is a very big role to play for, like, basic knowledge.

Speaker B: Tara Kolofsky, head of Technovation and also the motive force behind the AI Forward Alliance, which aims to unite leaders from tech, government, and education to equip young people and professionals with the skills and opportunities necessary to drive AI innovation. Something that our next guest, Dr. Antonio Weiss, the UK Prime Minister, Sakir Starmer’s former advisor on AI, is also a champion of. According to Dr. Weiss, it’s essential that we start to familiarize ourselves with AI, because if we don’t know how its plumbing works, we might be in a bit of a mess.

Speaker I: In order to have the most informed conversation at a big level, I mean, the thing that I would encourage everyone is, is to try these things, try these tools out, try and understand what they are, what they’re most importantly what their limitations are, how they work. So to give you an example why I think that’s so important is people often say, oh, you know, well, we’re gonna, I’ve got some books, I’m gonna train a model on my own files, okay? And I often say, and I try and point this out in the book, these models, ChatGPT, the early versions were trained on about 11 trillion words of data. You know, the King James I Bible has under a million words. If you have a few PDFs and a few emails in your inbox, you’re not going to be able to create your own bespoke model. You will still be entirely reliant on a frontier model. It’s just retrieving information. But I’m worried that we’re not maybe moving fast enough as a nation to adopt these. Well, not even to adopt the tools, but just to get familiar with them because other countries, and if you look at usage levels across the world, it tends to be younger countries. So India is a younger demographic, sort of skyrocketing use, they will adopt the tools faster. They will then work out their limitations quicker. They will then have the imagination to reinvent the way that organizations work quicker than us because they’ve just got to have more time to play around with them and realize what it’s all about. So I’d strongly encourage people to kind of start, if they haven’t done yet, start dipping their toes in the water of AI because it’s it’s not going anywhere.

Speaker B: That’s the fundamental point, though, isn’t it? We have to start to understand the infrastructure. We have to understand how the technology works. We have to experiment with it. We have to get familiar with it. We have to actually feel so familiar with it that we’re comfortable with it and comfortable with using it. And we’re nowhere near there at the moment. And we don’t seem to have that necessary underpinning. At the moment.

Speaker I: No, I agree. I worry about that, although I can’t remember who said this, it wasn’t me, but somebody once pointed out that most British, most people don’t know how a toilet flushes. There’s lots of things that we rely on which are really quite integral to our lives that we don’t understand. But now’s the opportunity, now’s the opportunity to try and change that because these things are quite different. And so it feels like this is the window to immerse ourselves in what they are.

Speaker B: There’s one thing, social media, for example, social media, we know that social media has created a lot of mental damage, particularly among younger generations. And we know that we’re waiting for that to actually emerge. I’ve seen people, CEOs, company directors of major firms using ChatGPT, and then waiting for the system to tell them what they’re thinking. That I find enormously alarming. I’ve seen that personally. And then I’ve also heard similar people say that the system thinks in the same way that they do. Now, that sort of anthropomorphization, confusion, whatever you want to call it, I can see that occurring over the next 5 years and that being very detrimental. I can also see overreliance of this from people who will start to become dependent on an AI system as a companion. I can see people saying that their AI companion understands them. It doesn’t. This is a computer system. I can see this terrible confusion. And that is something that we need to address.

Speaker I: Yes. And I mean, if you look at Harvard Business Review published their top 100 use cases So, you know, different ways that people are using generative AI in 2025. Top one is therapy and companionship and counseling. That is what people are actually using it for. So we, I spend a lot of time talking about some of the more commercial, professional, or service-based benefits, which could be in planning or financial planning, which actually is quite good that I think young people in particular seem to be interested in that now. But actually it’s around that social isolation element where it seems to have skyrocketed. And I think that is worrying. And it feels like we have to try and understand what the thing is and its risks and its benefits at turbo speed. So with smartphones, I, you know, smartphones, I would strongly say, and most other— most academics would say, is its own category of general purpose technology. I’m pretty sure historians will look back and say Wow, they took a long time to work out how to regulate and how to deal with that. You’ll have seen, I’m sure, and heard of in Australia the attempts to ban social media, or actually the legislation on that. France is aiming to ban for under-fives, and I suspect that is going to be a wave that’s going to grow more. I mean, just like regulation for everything lags behind the advent of new technologies, there will be things that we would, we would have regretted or will regret not doing sooner around generative AI. And again, I think the only way that we can understand those risks is by understanding what it is more and more. But yes, I mean, those concerns that you raise are concerns that I have.

Speaker B: Dr. Antonio Weiss, author of AI Demystified, a senior partner at the PSC, and to Keir Starmer’s former AI advisor on why we should know more about both AI and our toilet systems. Such concerns were consistent among everyone we interviewed, as was a need for a regulation that was designed to reinforce those concerns. To avoid the tsunami of mental health issues currently being stored up by mobile phone and social media use, we have to move fast to prevent even greater damage being done by AI use if we want to head off what some observers are calling an epidemic of AI zombie clones. It’s something that Berlin University’s Dr. Marvin Kubka, a psychologist and human AI researcher who specializes on how AI is impacting human performance and decision-making, says needs to be addressed both with regulations and by the technologists themselves with the systems.

Speaker J: I do not agree that in general using AI really has a negative impact on your cognition, or at least I wouldn’t say that this is definitely the case because we do not have enough evidence for such a claim yet.

Speaker B: But this is a problem, isn’t it? I mean, the problem that we have is by the time that we do have enough evidence, we have a serious problem. With social media, for example, people have already been saying Hang on, we’ve probably got some mental health problems that we’re going to have to be dealing with in the future with young children because of social media. Do we need to wait that long?

Speaker J: That is indeed a very good question and a great comparison. And I agree that we should probably have an eye out on the cognitive impact of AI. But I think instead of saying this would be a problem, I think we should vouch for solutions to tackle this problem right now. I think just banning ChatGPT or other AI tools from the classroom is not really helpful right now because students will use it anyway. It’s out there, they have access to it, and they will use it at some point. What I think is that for students and also the general public, we should teach something like AI literacy so that they really know how to deal with AI tools. So for example, the study that you referenced, it actually asked students who were not really familiar with ChatGPT and other AI tools to write an essay. And then they found that this essay was worse when they wrote it on their own later on, once they used ChatGPT. But I wouldn’t say that this is really a use case right now for AI, that you’re using it in a completely new use case and then you do the task on your own again and just that’s it. What I observed in my own classes, like I teach decision-making in university, and in my first year I had some students that also used ChatGPT, kind of similar to what the study reported. So they were not too critical, they used it, and they thought that they can outsource everything to it. Now, one year later, I taught the course again, and then suddenly the students were way more critical of the AI. So they said, okay, I get this recommendation, but it’s probably not true because AIs tend to hallucinate. So this is also why I said that I agree with the study’s interpretation of their data. But I do think that if people generally learn how to use AI tools correctly, so that they make mistakes, that they should cross-check the information and so on, then at least in my opinion, it’s not a big problem right now. It may even be helpful to outsource these lower cognition tasks similar to how we use GPS today, so that we can focus more on the street and navigating through the streets rather than having to consider where to go. I can imagine a future where we, for example, do not have to spellcheck the text that we write because ChatGPT or other AI tools just do it on our own. Does that mean that we are worse writers? I don’t think so, because the essence of writing will be more focused on the content, which ChatGPT is currently not too good at, and less on the formalities. So I think it’s just a changing world and we are experiencing this change rather quickly because it’s having such a high impact. So I think maybe it would be an idea moving forward to offer trainings or something. So that people really kind of learn and continuously engage in this kind of cognition that is needed to correctly use AI tools like ChatGPT and so on. So again, something like an AI literacy. And I also agree with you that we definitely need to train this also for the new use cases that are just arriving right now.

Speaker B: That’s the essential point. The essential point is that human in the loop. But that’s quite a difficult thing to accomplish. People have found in the American military, for example, that people are very, very hesitant to override, override missile systems because they don’t want to have the responsibility. They want to actually give that responsibility to computer systems. So teaching people to be the human in the loop, teaching them to have responsibility is almost going to call for a different sort of education.

Speaker J: Yeah, I agree. And I would also say for a new kind of jobs, like we, we do have something called human factors engineering right now. And this is a job type which is not too popular at the moment, but it is essentially designing human technology systems so that things like these do not happen. So that we really train people in the things that, that are important. That we assess when they do not want to override and when they may override system decisions and so on. So I think this may also be a new job opportunity in the future so that we really have people that kind of design these systems so that they educate people that they can and should override automated decisions, that they set up the whole system in a way that it makes it easy to override decisions. Like if you have to push 10 buttons and then confirm once again, people will not override it. But if they can just override it with one simple button click, then they are more likely to override it. And I think this kind of requires a systems engineering approach in the future that ensures that it’s really, that is really designed and set up in a way that leads to the best performance and the best possible impact.

Speaker B: That’s the issue, isn’t it? It’s our relationship with this technology, because we are the desire in the mechanism. It’s just not, it’s not just the human being in the loop. Legally, the software system doesn’t have any legal liability. So therefore, it’s a human being who has to make that decision. So they have to start making these informed decisions, they have to start forcing this relationship with the technology. And we are not seeing that at the moment.

Speaker J: Yeah, I agree. We, we are not seeing it and I think this is a huge problem because we are just kind of overwhelmed from, by ChatGPT because it was just released to the public. Now people are using it, people are making mistakes with it, people are learning from these mistakes. They kind of outsource all the different tasks to ChatGPT. And I also see some people that actually turn their backs on ChatGPT because they notice that it’s making them worse, that they stop thinking of the correct words, that they cannot handle the tasks that they have done before. So yeah, we’re kind of in a new space that is completely unregulated and not designed. So we do not have this correct system set up that makes it easy and good to interact with technology and AI right now. But we just have this, this, this tool that was released and now, now we’re in this situation. So I think there’s definitely a lot to do in the future to design these systems really well.

Speaker B: Berlin University’s Dr. Marvin Kopka on why we should redesign AI in people’s interests and not in the interests of Big Tech and its shareholders. Because in previous technological revolutions, the interests of those involved were transparent. Because, as Kopka says, technology is not wholly bad. A point that was succinctly made by Sergio Oliveira, the Chief Technology Officer for Design Rush, who we didn’t have time to interview, but contributed this. He said, I don’t think it’s fair to say AI is dumbing us down, especially when you consider about how people have always use technology. However, I understand that people have been anxious for a long time that new technologies may make people less smart. The printing press, calculators, and even the internet have all been accused of doing this. But in fact, these new technologies often help us learn more by handling simple tasks for us and allowing us to focus on the tougher ones. We often forget this small but important point, that AI isn’t making us less brilliant, it’s just changing the way we think. For example, AI is taking over jobs that used to need humans to use our hands, going through such large databases, organizing information, and doing simple calculations. This doesn’t stop your brain from thinking, but it does make room for higher-level thinking, creativity, and problem-solving. But to me, the true risk is becoming reliant. If we depend on AI too much to make decisions or solve problems, it could slow down our mental growth. Just like calculators didn’t make people worse at maths, AI won’t either. It will only make people less smart if we stop using our minds for those things. To go around in modern culture, you need to find the right balance. People should use AI as a tool, not as a replacement. And this is crucial for everyone to understand. For more on the ramifications of technology on society, go to our website, Future Intelligence, www.futureintelligence.co.uk, or sign up to our new TV platform, Tech TV. To get further insight into how technology is changing our lives and our world. You’ve been listening to Password on Resonance FM, produced by the long-suffering, patient, and wonderful Blue Buffery, and presented by me, CT1. Thanks for listening, and goodbye.

Speaker A: This program has been brought to you by Resonance FM. If you like what you heard, please support our work by making a donation at resonancefm.com/donate.