Brain Food for Pioneering Spirits: What is artificial intelligence actually doing to the world of work? In Room #856 of the Deep Talk Club, under our overarching theme “Pioneering Spirit with brAIn,” the group digs into AI hype versus AI reality — from clickbait job-loss statistics and the many kinds of intelligence, to trust, morality, automation history, and what education needs to look like next. Listen in on the examples that came up and the perspectives that emerged in dialogue…
Note: This session was recorded live in German as part of the Deep Talk Club, so the video above is German-language audio only. What follows is an English translation of the conversation, lightly edited for readability — the back-and-forth between the moderator and participants is preserved, but the original timestamps and speaker names have been left out.
A wonderful good morning to you all. After yesterday’s truly wonderful event with Dr. Johanna Dahm, the press conference that took place in Frankfurt — some of you were there in person too, and it was fantastic — I’m happy to start the morning today. I can tell you, I can still feel it in my bones. And yet, and precisely because of that, I’m totally buzzing this morning. And yes, I’ve brought you a link again, and I’ll explain everything in a moment. So I think today is going to be really exciting. I’m really looking forward to this exchange with you and to talking about AI and the world of work. So what actually is artificial intelligence, and what impact does it have on the world of work? That’s the topic we want to look at together today under our overarching theme, Pioneering Spirit with Brain. And similarly to today with AI, in the coming days we’ll also open up other aspects — whether that’s blockchain, other technologies, VR headsets, and so on. So we’re going to look at a wide range of new technologies here, in order to understand the connections between them, to realize what that actually means, how I might be able to use it for myself, and what it means not just on an individual level but also thought of communally, as a society. Where are the dangers? Where can we lend a hand? How can we still make use of it for ourselves? And what contribution can we make in this context? And what does all of this have to do with pioneering spirit, with brain? So it’s going to be very, very exciting, and I’m really thrilled about it.
And in this context, I’d like to start today’s AI topic with an announcement I’m very happy about. As I said, the Knowledge Base has been in preparation for you behind the scenes. We’ve put a lot of work into this over the past few weeks, built it up, and so on. It’s going to keep growing and being filled in, but the first draft has already been published. You’ll find it on my website, yaseminyazan.com — first and last name dot com. There’s a tab or menu item called Knowledge Base. There, from A to Z, at least one term has already been picked up under this umbrella, Pioneering Spirit with Brain. And a lot more will follow here. So this is the first draft — many of the cross-links aren’t in place yet, that’s something that will come step by step. I already announced that various highly qualified experts and scientists will be contributing here too, delivering their own pieces. Those conversations are happening right now. And to keep the quality high, I’ve decided that these will be people who are directly connected to academia — typically people who hold a doctorate, who teach at universities, and so on. So a lot is going to happen here in the coming days and weeks. I’m looking forward to us being able to publish this first draft, so to speak, rather than saying, well, speed of implementation is an issue here, so we wait for weeks until some contributions arrive. I decided instead to simply pick up material that already exists, add some new pieces, and publish it accordingly. That means the entries range from very short to very long, all sorts of things, just so that we at least have something under every letter. So don’t be surprised that it’s not fully comprehensive yet — it’s going to keep growing, and I’m really looking forward to that.
And why am I mentioning this right now, today? Because it fits our topic of artificial intelligence — there’s a piece posted there too, which was actually written some years ago, so it’s not new either. It existed in English before, and it’s now been translated specifically for the Knowledge Base, in order to cover artificial intelligence with this piece as well. And with that I’d like to start with an impulse, so to speak, for anyone who maybe doesn’t yet know exactly what artificial intelligence actually is. Roughly summarized, it’s about the ability of a computer, or a computer-assisted robot, to carry out various activities in a similar way to how intelligent living beings do. That’s the goal behind it. And that’s why, in contrast to natural intelligence, we use the term “artificial intelligence.” A great deal of it is, at bottom, an attempt to mirror — roughly, or even one-to-one — the structures or processes we know from, or attribute to, our own brain. And that’s why in this context we speak of an artificial intelligence, in contrast to the classic human intelligence we know as such. Artificial intelligence therefore stands for machine intelligence that uses information gained from data to solve complex tasks, advance research, and better predict events and human behavior. And as a result, more and more intelligent machines are emerging that can work and react like humans. This is a groundbreaking technology we’re talking about here, one that, in combination with other technologies, is increasingly changing our world. That’s not a new insight — we’ve said that again and again over the past years. The only difference is that there was a development just recently, a few months ago, where ChatGPT, an AI-powered tool, was made accessible to the market. And that has led to a situation where, even though in everyday life we’re already consciously or unconsciously working with AI a great deal, one person or another has now realized: oh, what’s actually happening here? In other words, what was previously invisible to many has now become visible, which has led to a great many discussions around AI — dangers, worries, and everything that’s being raised there. It’s important to underline very, very strongly again: this is not a new technology. This is technology that people have been working on for a very, very long time. It’s only through this market penetration via ChatGPT, through the free access, and through our awareness being sharpened — sharpened for many people — that we now have a situation where suddenly people are discussing something that wasn’t so prominent in the media before. So that, perhaps, as an aside first.
And yes, ChatGPT is just one example. As I said, AI tools are used and deployed in many, many contexts. And ultimately, this is a tool like many other tools we already know. You know this from everyday work — whether that’s emails or various software programs we use. Ultimately it’s about achieving an increase in efficiency by delivering certain work results faster. And the groundbreaking thing about this system is the deep learning running in the background. In other words, this machine learning that’s taking place here means that information that’s fed in can, so to speak, be recognized and recombined in new ways. And that produces yet another exponential increase in the potential for efficiency gains that we have here. And because our brain thinks and imagines things in a very, very strongly linear way, we sometimes really do have difficulty grasping just how much of an exponential wave is rolling toward us over the coming weeks, months, and years. And of course that always has to be looked at from two sides. The concerns are not unfounded either. You can look more closely at what those concerns actually are, and which ones are justified or unjustified — that’s something to discuss on the ethical level, so to speak, or on the societal level. And the other aspect, which is actually what we want to look at more closely today, is this one: where can I apply all of this? What impact does it have on the world of work, and where are we actually heading? Let’s do a bit of brain-juggling here together and just take a look.
And in that context, alongside many other fields of application that you can also bring in, I’d like to build a bridge to the link I brought along today, and also build a bridge to yesterday’s event — because yesterday’s event with Dr. Johanna Dahm, with the fantastic co-authors, with editors, wonderful guests like Uwe Bingel, Julian Backhaus, Mick Knauf, Roland Tisci, and many, many more, was truly an inspiration, a pure fireworks display. It was about decisions, and also about the role decisions play, how you achieve success through decisions. And I think that fits wonderfully with our topic, because AI has various areas where we can use it, and one of those areas is precisely the context of book writing and authorship. I’ve already placed this in a few spots — you know that the “Magic Hack for the Optimal Use of ChatGPT — Create Your Own Specialist Book with the Help of AI” was created exactly as a pilot, so to speak, in order to demonstrate to people who haven’t had much to do with AI so far what’s actually possible with it. Of course you can look at the book and say, well, I could write that much better myself, or whatever. But the whole point was simply to show what’s theoretically possible with this kind of efficiency gain. Because this book, within this pilot that was launched to demonstrate exactly that, was created within 24 hours. And I think that’s a wow effect, because it simply shows what’s possible.
At the same time, energized by yesterday’s event, I got up relatively early again this morning — after basically collapsing pretty quickly yesterday after the event because I was tired — and I started reflecting again. And one thing struck me in the conversations. Even though there were such wonderful people there who, on the one hand, show just how important decisions are in order to move toward success, whatever that means for you, and what it takes — at the same time, we all keep falling into the same trap, and I don’t exclude myself from that. And that trap is saying, yes, the next thing on the list is this or that, but there’s still a bit of time for that. And interestingly, a lot of people approached me yesterday, whether about mentoring or about book projects, and book projects were at the forefront. That is, many of the authors who took part said, yes, I know the next book project is coming up, I’ll see when I start with it, I’m not going to look at it today, maybe tomorrow, so to speak. And that’s exactly the human trap we fall into in the context of decisions. And especially in the context of AI, this now has a massive impact, and I want to raise awareness of that — not to put things off, but to really move into implementation, to make decisions, to get going. Because the moment I start moving, the rest follows too. It’s only when I don’t even start, and just keep postponing, that not much happens. And I asked someone that question yesterday too: the question isn’t just what price do I pay for doing something, but also, what price do I pay for not having started, for not having set off. And I asked that question in a very personal context yesterday, in a personal conversation, with someone who has been mulling something over for two years and still hasn’t started. I don’t want to name any names here, but it was really very concrete. And if we look back, two years have passed, and a lot has happened in the meantime, but what this person originally intended hasn’t progressed at all.
And that’s why I’ve brought this link along again today, and I want to place it here too, especially in the context of AI, because this is one of the fields of application we have here, one that I work with on a daily basis myself — namely, the Bestselling Authors Club membership. The Bestselling Authors Club is, essentially, for everyone who’s toying with the idea of publishing a book — it’s the all-round membership for everyone who’s willing to go the extra mile. That means it’s step-by-step support from the idea all the way to implementation, whether for ebook, paperback, hardcover, audiobook, whatever you want to publish, including marketing strategies and measures, both in preparation and with a view to bestseller status, and then also the marketing afterward — because a lot of people forget that it’s not finished, so to speak, once it’s published, but that’s actually when things really get going, in terms of deployment and distribution with maximum added value. And on top of that there’s now something new, which I’m very happy to announce and promote here today — namely, AI-powered tools have been added. That means it’s not just about how to write, what matters for marketing, and so on, but also about pointing out the corresponding AI-powered tools we can use for ourselves in the German-speaking market, in order to achieve a thousandfold acceleration in implementing the project. So, take a look — anyone who’s genuinely toying with the idea, take the opportunity. It runs for a year, and it’s not just some online courses or videos you can watch, but also comes with regular monthly live online support, where — and I think this is really important — you have the opportunity to formulate questions, so that next time you also get asked again: so where do you actually stand? Have you gotten three steps further, or has it just stayed at the planning stage again? And I think it’s exactly this accountability, this follow-through, that’s also something we need, something that moves us forward here. That’s why I wanted to bring this in as the first example here — artificial intelligence, one of the areas we use and apply on a daily basis, here in the context of book authorship. And how can I actually use and apply these tools? What can I do with them? And that’s far more than just writing the text — by the way, what these tools offer in terms of possibilities that I can use to become far more effective at achieving visibility, at shaping marketing accordingly. From images that can be designed, to text passages, to posts, to social media marketing, and so on — so the whole range of things I can use not just in the context of book authorship, but also everything to do with marketing in other contexts for my company and for my business. So that’s one example of where AI is being used today and is massively unlocking, unfolding, and accelerating a thousandfold the potential in our projects and in our speed. Of course, on the other hand we mustn’t forget deceleration either, I’d like to underline that too — because if I have the possibility to accelerate, then on the other side I also have the possibility to decelerate again, through the free space that’s been created for me, and to make use of that too. And I think that’s a wonderful possibility as well.
So, that’s one example. There are many more — I could bring in a lot more examples — but I’d really like to move out of this monologue mode now and step together into a shared exchange, a shared reflection. What are the areas in the world of work where you’re maybe already using AI today, where you know that we’re deploying AI, that AI is running in the background, so that we can raise awareness of that and then look together at what kind of impact this is actually going to have on our world of work. And with that I’d like to warmly welcome Marc — a wonderful good morning. I hope you’re feeling fit.
Good morning. Yes, I’m fit. I had the great fortune of also being at that event. I also only got four hours of sleep, but it was fantastic, and there were so many topics, we can say that too. I saw Yasemin there of course, and I picked something up. And a lot of the topics we’ve discussed here were also raised there. So we’re not the only ones dealing with these subjects. Before I say something concrete about the topic, I’d like to share an insight with you that’s important before I get to my statement. It became clear to me again yesterday: we source information. In companies we get information, and we also get information from the press. And what I keep noticing, or keep having to apply to myself, is that we should check who’s putting this information out, what’s the truth content of this information, what purpose does the person putting the information out have — is it purely for the added value? And that we should always keep reminding ourselves — and I’ll give an example of this in a moment — that when this information comes from the press, these aren’t bad people, but negative headlines earn us money. So the press informs us, but in order to sell the newspaper, we need to be startled. And I can explain that with a small example. When you look into the topic of AI, you find several options on the internet. The first option that immediately catches your eye is: watch out, 800 million jobs will be affected by AI. Whoa, I think, 800 million jobs — wow, that’s quite a lot. That immediately stirs up fear. Or, oh boy, what’s coming our way? But if you keep researching, you also find other reports that say: 80 percent of workers will be supported, or are in jobs where at least one task could be completed faster with the help of AI. If you compare these two texts side by side — on the one hand a task can be done faster, or on the other hand 800 million jobs are affected — that produces completely different feelings. That, first of all, as an introduction.
And before I get to which professions will disappear, it’s also really, really important to me to start there — because yesterday too, we heard a fantastic talk. I hope the man comes to an authors’ breakfast sometime — Mick Knauf, a German financial journalist who covered the stock market, the topic of finance, so fantastically. And if someone had said to me, Marc, don’t you feel like coming to an event with me tomorrow evening about stocks or finance, and someone from a bank is speaking — I would have said, are you crazy? No, with all due respect, not really — I’d rather sit outside on my terrace and look at trees. And this person, this Knauf, covered the topic with such incredible enthusiasm, with such simplification of how the markets work — simply, and you don’t even need a degree for this — that if we just switch on our normal brain now and then, we get a lot of answers from that. And now, what I hinted at, and then I’m done: what does AI do, before we list the professions? AI gathers information, AI will process information, and AI will also pass on basic information. And once you leave that as it is, it automatically occurs to you which professions could be affected by that — gathering, processing, and passing on information. And I’m not going to go into the specific professions right now, and I’m glad Jürgen has joined us too.
Yes, wonderful, thank you so much. And I made the contact directly today, just now, via LinkedIn. And he’d basically already agreed yesterday, at least verbally. Now let’s see if we can establish the contact and invite him to an authors’ breakfast or an expert talk or whatever it ends up being. So of course yesterday we were already thinking ahead and making sure that the various contacts made there can also be used and made accessible for the community. So I’m curious — let’s keep our fingers crossed. Jürgen, a wonderful good morning. What do you say about the topic?
A wonderful good morning, dear Yasemin. Good morning, dear Marc, and everyone in the room. I hope I can pull all this together — you set off so many trigger points in me with your opening, and you too, dear Marc, with your approach to the topic via AI. I’ll approach it with my statement from the other side, namely from the side of intelligence. And I’ll place this contribution under “for what purpose” — in the sense of reducing complexity. And that was also a trigger, dear Yasemin, from you — that it’s, in inverted commas, human to very often say, especially in project work, that now isn’t the time for that, or the time for that hasn’t come yet. My experience is that a major factor influencing that statement is unmanageable complexity. And so, hopefully visibly, here’s my train of thought: we’re agreed that this is a huge topic — I’d like to voice the wish that it should actually be an overarching theme in its own right, for me, because it’s immense, both in terms of its effects and its complexity. But I’d like to reduce the complexity a bit from the outset, if I may remind you of another overarching theme we had — education, as so often. You’ll remember we talked about learning, about socialization, about knowledge, about intelligence and reflection. And together we arrived at the conclusion that there is no all-encompassing, universally valid definition of intelligence. Why? I would have one, but I’ll leave that aside for now. Why? Because there is no such thing as “intelligence” as a single thing. We agreed, maybe you remember, that we mostly identify or see intelligence as mathematical-logical competence. But there’s also emotional intelligence, which enables people to work in, say, caregiving professions, or personal-leadership intelligence, which enables people, further down the line, to take on leadership positions, or kinesthetic intelligence, which predisposes people to become athletes or dancers, and so on — you could add plenty more. But the important thing is: there are multiple kinds of intelligence. And my question now, looking at this topic, is: what influence does AI, artificial intelligence, have on the world of work? Which kind of intelligence is actually, or primarily, the one that influences the work environment we have?
Just a personal opinion of mine: I do see AI as placeable in the mathematical-logical context, and probably in the main field of deployment over the coming years. But in the kinesthetic realm — this might sound ridiculous — athletes or performers on stage, that’s not going to happen. And that’s exactly it, dear Marc, and you spoke straight from my soul, exactly this unreflected way of immediately slapping a hype exclamation mark on statements like “this will change the entire world of work.” No way — I won’t claim that. And I hope I’ve given my reasoning for that, because intelligence has to be seen through very different facets. And regarding how AI will be shaped and how it will develop over time, I do have some partly diametrically opposed opinions on that too, but I’d like to set that aside for now, whether we even get to that topic. So that’s my first input on this. Reduce it to what’s actually possible and considered probable from today’s perspective. I think that helps. Thank you.
Yes, thank you so much, and let me follow up briefly here to make sure we’ve understood you correctly. What exactly, in your view, does not change in the world of work because of AI? Maybe you could give a few very concrete examples.
Well, dear Yasemin, I’ve read a great deal about this, but I won’t presume to claim the label of expert for myself. But one starting point for me, for thinking this through — and you’ll notice I’m trying to hold back a bit — is classification by routine, where I do see, and please bear my earlier remarks in mind: a mathematically-logically based AI will certainly provide support for a lot of routine work, dear Marc — well, first a distinction between support and replacement. But routine has two aspects. There’s mental routine, and there’s physical routine, like brushing your teeth every day. That already opens up a first slimming-down of complexity for me, in that certain mental routines can be replaced, like what we’re seeing right now — collecting data. I have a very good example in my own family circle, I won’t make it too long, because — unfortunately the train has just arrived — but he’s a partner at a tax firm, and the same applies to lawyers. The core work needed to arrive at an individual decision — is this a tax offense, yes or no, or is it a crime or a minor administrative offense, and so on — involves an immense amount of data collection beforehand. And there I see quite substantial, and I’ll put a question mark on this, replacement potential or complementary potential — that alone could be its own topic for me. But that’s one point for me. And to sum up quite clearly again: physical routines are off the table for me in that sense — though I could give an example, I’ll spare you that now, there are considerations along those lines too. I hope that answers your question.
Yes, thank you so much. And I’m really curious now what Layla, and then maybe Marc too, will add to this. Here too, I think we should be mindful not to think in either-or terms — because if we pick up these terms again, replacement potential or complementary potential, then we get chain reactions being triggered. That means, the moment I deploy it as, let’s say, at least complementary potential, something follows from that — and we’re already observing this today in the labor market, it’s not new, it’s already started. The moment I have the ability to supplement an individual’s potential — and this wasn’t new with AI either, it was already the case with other tools before — I have an effect, namely, for example, that a company thinks, well, in the future maybe I won’t need 100 employees anymore, just 50, who then do the remaining work, because a complementary potential has been unlocked, so to speak, and so, viewed as a chain reaction, we end up moving into a replacement potential that can’t be avoided. However, to take some of the negative emotion back out of this — at the same time, if we look at it historically, it’s not the case that — sure, tasks shift, jobs shift, and so on, or wherever demand on the market happens to be. But that doesn’t necessarily mean that, as a chain reaction, it has to trigger mass unemployment or anything like that. And here we’re welcome to discuss again what the effects might actually be. But at the very least, there will be different tasks in the future, different areas of focus. And it will also mean that companies really will, because they have the option, reduce the number of employees — and that’s already happening today, thousands of employees are being laid off, for example because banks have discovered, well, we can use an ATM, we don’t need a teller, we reduce staff through online banking — and we’re not even talking about AI there yet, just general automation that’s taken place in the past, which means the mass of people in that particular activity, as it existed before, is no longer needed in that form. But it might be needed elsewhere, and that’s exciting to look at too. Jürgen, you wanted to add something.
Thank you, sorry, dear Layla, that I’m picking this up again — you touched on a second point I mentioned earlier, I’ll spare myself repeating it — others that might be, or are suited to, reducing the complexity of this topic. And the ATM is, for example, one such position — it’s my specific field, banking — so let me get to a point that’s representative for me. Of course you can — the ATM spits out money for you, but you can’t look the ATM in the face, in the eye, and ask, can I still trust the euro? That’s the issue for me, and I’ll put it this way: you can’t look a machine in the eye, and if we only had machines left, we’d be giving up a quality, or rather it simply wouldn’t exist anymore — namely the trust that’s based on personal relationships, including in business life. And the second key point, and again a very strong plea here: not everything that’s possible is also economically sensible. That means, as brutal as it sounds, a company will always weigh up what’s cheaper — the machine, the artificial intelligence, or the human. And the more people I have — because AI might free up a lot of people — the cheaper the resource “human” becomes, and the more likely I am to decide, before I bring a highly complex system of algorithms and the necessary data infrastructure into the house, I’d rather just have it done by hand. So thanks again for letting me bring in that point.
Yes, gladly, and there are certainly some really exciting aspects to discuss in this context too. I’d like to bring in Layla and Sabine now, and then we can certainly draw cross-connections again. Layla, a wonderful good morning.
A wonderful good morning. Yes, so you’ve all mentioned a great many thoughts — I’m trying to figure out where to pick up. I’ll start by saying, well, I hope you’re not thinking about what’s cheaper, because I learned in business school that “cheap” means poor quality, and my teacher always told me to say “affordable” instead. And here too, I think, to find the frame again, it’s often about — I think we have a lot of narratives too, things we’ve seen in shows, like, oh, what’s happening with AI? Let me think about this practically, to give an example — how can AI make the world of work easier for me? I listened very closely, especially regarding the book — I found that very interesting. I’m already using it in my work, having it help me with lesson plans, bringing in ideas. Of course some areas will be replaced, and it was already mentioned that the ATM is one such case. I know computer scientists I’ve talked to who already work with GPT-4. I also asked them if they’re afraid of being replaced. They said no, because AI can’t do this or that. Of course you could say it’s developing very fast. If I think about my own profession — I’m an ethics teacher — let me think, Yasemin, which source I’m allowed to name, but I think Dreisatz is fine, and Skobel is a recognized philosopher — an AI has no morality. And it also doesn’t have eyes yet, and it doesn’t have morality yet. That’s the trolley experiment you can run through, thinking about what, what can I — not what the AI can’t do, where can’t I be replaced, but also, where can it support me? And that’s actually how I’d like to think about it. And yes, morality simply isn’t built into it. And you can’t blame it for that either — it’s not something you can hold against it negatively, “you have no morality.” No, but that’s something we humans can certainly have. And yes, emotional intelligence too — I don’t actually know if that only applies to caregiving professions, but also to social professions. Salespeople need emotional intelligence too. Where do I apply it? And I think those are points that are very positive, where I can say, as support. And yes, we already have this, but it’s also a common complaint about hotlines — we’re already talking to AIs so often, especially when we have an assistant, some kind of chat function. And here, I think, the human being is an addition. The AI should be the addition, not us. And yes, business owners think that way — that’s true. I hope they also think in terms of value, not just cheapness. What’s the priority? And yes, as I said, I’m trying to make my own work easier this way, and I’d recommend it to people too. But I do think there are areas — if we’re talking about morality — for example, Snapchat now has an AI that children can use. And yes, I was quite alarmed when the AI says, send me a photo, or don’t tell your parents. We can’t blame the AI for that, but we could blame the people who built it. Why are you using children as test subjects for this AI? And of course you could say the legal situation left a loophole here. But well, I’m not the ethics council of the world, but yes, I just wanted to raise that for consideration. But I find it interesting, great, and I’d like to make use of the support it offers. Thanks for listening.
Yes, thank you so much. And to pick up on two points, to take up what you said, and then switch over to Sabine — first, let me throw in this impulse: actually, AI does have a morality — quite provocatively, quite deliberately phrased that way — namely the morality that we, as humans, assign to it. That means, ultimately, it’s a programmed system, and depending on what we define as morality and feed into the machine, it will act accordingly afterward. So it does have a morality — not one it’s worked out for itself, perhaps, but one that we program into it. And I’d like to follow up on that right away, namely on the many concerns being discussed here. Very often what I hear is something like, I don’t know, it takes on a life of its own, there’s discrimination, or whatever — to raise awareness here too: the problem, and at the same time the solution, is the human being, not the robot. That means — what does AI draw on, then? The data that we feed into it. If the data we feed in only uses masculine forms, if discrimination is already happening today in our texts, in job postings, in the selection processes we run, and so on — then we’re at the human problem. Unconscious bias — it’s documented everywhere. That’s the information being fed in here. So I shouldn’t be surprised if AI then discriminates too, picks up unconscious bias, and so on. And the solution, I’d like to underline this too, is likewise the human being. Because if we manage to program different things here, to raise our own awareness, to see the dangers, whatever that takes — whether regulations, an ethics council discussing it, or whatever — then we can also manage to build in the morality in such a way that it represents potential rather than danger. That’s one thing. And the other thing I also thought was great is that you brought up chatbots as an example too. That shows that we’re already using this in so many places today without necessarily being aware of it — whether that’s ATMs in the sense of automation that we deploy, or online banking in the sense of automation that we use and deploy. Those aren’t classic AI-powered tools, but the automation alone that’s taken place in recent years shows that this is also what a lot of people actually want. It’s succeeding partly because, just as an example, people aren’t willing to pay bank fees just so someone can sit at a counter somewhere. Every one of us looks around for a checking account that’s free of charge.
And at the same time we then say, no, we’re somehow not okay with that, from a societal perspective on the problems — but because the benefit is there on the market, and because it’s convenient for us to use online banking, in the end we still choose online banking. And it’s somewhat similar with all these AI-powered tools — there’s no way around the fact that this will assert itself here in terms of market penetration. On the point of “cheap,” in inverted commas, or poor quality, just briefly again: even if it might make sense from a certain way of thinking to say, well, if the resource “human” becomes cheaper, then as a company I could also consider whether I now prefer the resource “human” over AI — I’d counter that as an impulse by asking, well, but who then has less potential for error? If we look at what data and at what speed results are being produced here — I’m not talking now about the parts that only a human can handle and that AI can’t fulfill, but about the things I can hand over to AI — then, in terms of the result that comes out at a certain speed, a human won’t be able to match that, full stop, in my view. Take medicine as an example: if I have a doctor who needs to make a diagnosis, they can only draw on their own store of experience, so to speak, and their own knowledge, while AI here draws on the knowledge and experience of all sorts of doctors in that field, and can recognize things that even a very good doctor today can no longer identify, let alone at this speed. So we really would need to think again in terms of error potential and things like that, and not just in terms of what’s cheaper, in inverted commas. At the same time, though, of course the problem arises that I have to consider — is it only ever about cheap? What does that do from a business economics, from a national economics perspective, and so on? That’s the next aspect. Because if, theoretically, purely hypothetically, we didn’t assume that new tasks would arise but that mass unemployment would really result — just to play through that option — that would mean people no longer have purchasing power. And that in turn, viewed as a chain reaction, has effects on businesses too. That means we end up in a cycle here that we mustn’t view as great either — it really makes sense to think in chain reactions here and consider what the solutions might be. And solutions could, just as an example, include considering whether, if companies employ fewer staff in the future, there might be taxes for companies that use AI — because that money has to come from somewhere to keep this system on its feet. So there are many, many ways to think about this, many directions we could go. And that’s exactly why this is a space for juggling ideas, so to speak — nothing is off the table, and simply opening up various options and thinking them through is, I think, the best thing we can do in this context, just to sensitize ourselves to the directions this could all go.
Sabine, a wonderful good morning, thank you so much for your patience.
Wow, good morning everyone. I don’t know if I can still gather up all the points. First of all, I certainly belong among the “magicians,” I’ll out myself here. Secondly, on the comparison with the ATM — I wanted to give a small example. I work with a Volksbank. They’re already three steps ahead. They’re in a rural area and don’t just have an ATM — when you walk in… I’ve unfortunately forgotten the name, but a screen comes up, and then a young lady makes a decision, and she has a conversation with you. And they tested this and it was very well received. So routine counter business is now being handled by this “lady,” and if she runs into trouble at some point, she simply says, I’ll just hand you over to my colleague. So we’re not that far away from that anywhere anymore. What I’d also like to raise for consideration is simply this: I’ve been trying for about four weeks now to personally — I’m already legitimized, my husband too, for our minor son — you can become a postal agent — at the post office, both post offices in Mannheim have unfortunately closed. DHL doesn’t handle it anymore. I’ve now spoken to everyone I know, they say, we’ll send you a new link for video verification — doesn’t work. What I just want to say with that is: the missing link is precisely that you’re already legitimized. That means, the moment I roll something like this out, I obviously need to make sure behind the scenes that it actually works. But that has nothing to do with AI, that’s about the whole technical setup — not skipping the first step in favor of the second. What else did I want to say… oh right, ChatGPT — no, sorry, what’s it called, what Layla mentioned earlier, that tool that, well, my kids also use, only I use it with the kids. And there I’m a bit on Layla’s side. Of course there’s some person behind it programming it, saying, send me the photo, send me that. But that’s a bit of the danger, isn’t it. I fully agree with Yasemin here — AI can’t do anything we don’t feed it and don’t implement in it. But the dark web is active at that very same moment too. And who makes sure that the right thing gets passed on to it and not the wrong thing? That’s something I’d still like to know. Otherwise I had a fourth point, but I’ve unfortunately forgotten it. Thank you.
It’ll come back to you, I’m sure. Yes, thank you, thank you so much. And with that I’d like to give the floor again to really make some cross-connections. Marc, you raised your hand, and then Jürgen, please.
Yes, I’d like to jump on that right away, dear Sabine. Just recently I read, for exactly this reason, who controls this — Italy has currently blocked GPT because they said it’s not at all clear what’s being done with these data requests, how the data is processed, where it ends up. I thought about that for a moment and then thought, okay, what about that other big search engine we use every day — what kind of data does that collect? So it’ll only be a matter of time before that resolves itself too. Then I’d also like to say, results-oriented — we talked about people, and we should be clear that if we have a company that works results-oriented — in a job interview, 60 to 80 percent comes down to likability. That’s not even the references, it’s likability. And in the discussion — this isn’t a value judgment now — it’s actually kind of crazy that bosses and companies hire people based on likability rather than their abilities. And the second thing is responsibility — we haven’t used that term yet — exactly, because we’re discussing all these things now, what’s coming our way, there needs to be responsibility for how we handle data, and, and, and, who controls the data, what effects does that have. And on the topic of “cheap,” as a business owner I have to say — I won’t name the car brand — there are of course also entrepreneurs who take completely different steps, who simply say, do we even want the cheap market at all? No, the market of the super-rich is far more interesting, and in the automotive sector there’s specifically one car brand that said, for us, cars only become interesting from 130, 150 thousand euros, and there are simply people in Saudi Arabia and Dubai who buy ten of them at once in different colors. That’s how they make their money. So this mass and cheap market is exactly what some companies we know are currently leaving, because they say, we don’t make any money with this mass market anymore, we’re focusing on the premium market. And of course that also has the advantage that, if you’re in the premium market, the quality — and there we are again with quality — has to be absolutely top, otherwise you’re obviously not going to sell a car for 120 or even 150 thousand euros. And that’s totally fascinating, like with every topic — no matter how we look at it, each of us finds their own answers, but I agree with what Jürgen also said, not to act out of panic but simply to approach it with a completely objective examination, with fact-checking, and if you really engage with this AI topic, the majority of the studies you can currently find online assume that it’s more likely to lead to a restructuring rather than mass layoffs — although of course it’s also possible that certain jobs will indeed disappear. A restructuring can, of course, and this was Yasemin’s point, also mean that for certain people in a company, jobs they used to do no longer exist, that they move into other jobs, and then something happens that we’ve already had as a topic before — that for those other jobs, you might get paid less if you don’t have the education for it. We’ve had that topic too. Education is going to become a really important topic, because if you want to work in AI, whether for programming or manufacturing, you need a certain level of education. That’s not going to work if you can’t read. And there you see, we’re back at the topic of complexity, and probably one room isn’t enough for this topic — we could easily do this room five to ten times over on AI alone.
Yes, certainly. And before I get back to Jürgen, then Layla, then Sabine, I’d like to briefly pick up on this concept of education, because that was also one of the topics at the authors’ breakfast with Dr. Franz Hütter, by the way — very worthwhile to listen to, because assuming that no jobs will be cut here would simply be an illusion. That would be a certain naivety. So jobs are definitely going to be cut here, full stop. That’s coming, and it’s already underway. Various companies today already — I won’t name names because I’m not sure it’s been communicated yet — but one large company is currently in the process of cutting 10,000 jobs, among other reasons because of AI use, and so on. And many other companies, including in the mid-sized business sector by the way, so also small and medium-sized companies that have already looked into this, are hiring less or cutting positions. So it’s already rolling, and a lot of people haven’t woken up to it yet. That means a whole lot is still going to happen here. But I don’t want to present this in a negative light — and that’s where we’re at this point about education. It’s becoming clear that we need certain overarching skills, beyond pure knowledge, so to speak, which is what may have partly defined us up to now, independent also of emotional competencies and so on. And you can look at that too — in the therapeutic field, for instance, it’s incredible what’s being developed here that’s groundbreaking with AI, that therapeutic settings are now being created that are run by AI, because, for example, there aren’t enough therapists — or a crisis helpline or something like that, AI-supported, that’s just amazing, everything that’s coming here. And rather than assuming, in the sense of, that won’t work — if I have the choice between waiting six months or a year for a therapist, or taking AI as a therapy option, well, what’s better for someone who might be at risk of suicide? So looking at it from that perspective too — there are so many opportunities we can seize here, both as a company and viewed individually, and also viewed societally. Because this is also a political discussion we’re having right now, also with skilled workers from abroad and so on, which of course also brings advantages and disadvantages. And at the same time, seeing what role education suddenly plays again — namely, we’re at the aspect of key qualifications, of overarching qualification. This needs to be supported very, very quickly and very intensively. That means the demands to deal with complexity, and so on, are becoming more important than ever. They were already important, but now especially with uncertainties, areas of tension, dilemmas, and so on — that’s becoming very, very important. And just conveying knowledge in a lot of institutions doesn’t help us at all here. The pure qualification mindset simply needs to be shelved. And rethinking this, and above all moving into action — I think that’s an incredibly important aspect for establishing, or re-establishing, self-determination, self-organization, and self-efficacy in the long term. Jürgen.
You unfortunately can’t see me grinning right now, but this is grist for the mill — especially, of course, always the quality of your synthesis, but today also the length of the remarks, how complex this topic is, and yes, dear Marc, I already positioned this at the beginning — this topic would really deserve to be its own overarching theme. And thank you for your input regarding what I raised — looking a machine in the eyes, trust is lost — I see that as a mirror image for leadership quality, or rather for the topic of responsibility. A machine taking on responsibility for correctness, for truth, for the ability to learn — and that brings me to the next topic, which I think would also do a lot of good for what you might call complexity hygiene: getting clear on what the actual difference is between automation, machine learning, and artificial intelligence — with a view to the final development stage, in the direction of mapping neural networks the way our brain has them. So that too, and I keep coming back to this point, is actually far too complex a topic not to make it its own overarching theme. And one last light-hearted remark at the very end, because apparently there’s some resistance to the term “cheap” — in economics, or in business, there apparently is a distinction, which is why there are the two terms “cheap” and “affordable” — there’s quite a strong distinction there. I won’t spare you the details, but for anyone who wants to know, I’ll gladly point you to it. Thank you very much.
Yes, wonderful, thank you, thank you so much. Layla.
Yes, I feel a bit like the devil’s advocate today, and I’d first like to thank you for that, Yasemin, for your impulse — because it’s not quite so — the impulse was, AI has no morality. And I think that’s good, because language is very important and it creates realities too. No — it has the morality of the programmer. And we see that, and these are discussions that already come up around autonomous driving. And you’re welcome to look this up — it’s called “The Moral Machine,” from Harvard University. Try it out and take a look — you have to make decisions there, and they’re very complex, and they’re hard for all of us. What I want to say is, I’m very positively inclined toward AI, that’s true — and yet it’s also true that I say, it has the morality of the programmer, and there are economic interests behind that. That’s simply what strikes me. And that strikes me with Snapchat, when children are told, send me a photo, and then the clock gets praised, or, which headphones are those, or, don’t tell your parents you talked to me — and an AI doesn’t think anything of that, or very little, or there are economic interests behind it, because we don’t know how it’s programmed, or who programmed it, or why it’s like that. And this isn’t meant as doom-mongering, we shouldn’t do this — rather it’s more of a warning directed at humans, and at economic interests, and we see that in business too, because we have no global ethics council, we have no universally valid morality. And that’s a point that strikes me, and that can be a stumbling block. Nevertheless, we can neither close ourselves off to this development, nor should we — we should make use of it. So I’d like to say thank you again for the impulses and the discussion, and I think it’s important — how do we use it? And that reminds me right now of my health insurance provider, whom I called — and if I’d been talking to a bot instead, that wouldn’t have been so nice, because I had an individual case, related to my chronic illness, and my caseworker had exactly the same illness and could really relate to it, and saw the individual case and did everything to make sure it was handled accordingly. And I don’t want to say that AI doesn’t do that — I just say, I don’t know if it’s always handled that universally. And that’s why I’m quite glad I had a caseworker who could relate to it. And that’s really the only thing. I want to use it, I think it’s great, it doesn’t matter whether we call it affordable or cheap — I just think it’s important that we take a really close look at it, and that we don’t forget: the problem isn’t AI, the problem is the human being, and maybe the problem is economic interest. And if we get that right, and feed it in the right way, this can be great. But if we only see the economic interests behind it and it gets fed in that way, then it’s possible that difficulties arise, situations arise that are difficult. But I use it too, I like using it, and I’d like to point to the link again, because I think the book idea is great, and yes, I’d be happy if that produces great books and great ideas too. And yes, thanks for listening.
Yes, thank you, thank you so much. And before I move on to Sabine shortly, and then we go into our closing round for today, I’d like to pick up two more aspects. “The Moral Machine” and “Harvard University” were keywords that came up in this context. I’d like to tie that back to scenario thinking, which we also had at the authors’ breakfast with Dr. Franz Hütter, by the way — really worth listening to for anyone who hasn’t yet, or if I was live there myself, worth listening to again — because it also raised awareness that we need more training in scenario thinking, and that this is one of the overarching skills we really need here, whether that’s working with AI, or how programming works, or other things, or other contexts — it comes down to scenario thinking. I really want to underline that here, and if you haven’t listened to it yet, definitely tune into that episode too. By the way, we’re also expecting Dr. Franz Hütter again in the context of VR, also for an expert talk — I’m really looking forward to that too, that’s next week I believe, if I’m not mistaken. So it’s really worth looking at what developments are happening there — that’s one thing.
And the second thing I wanted to pick up on again was that whole example, the one with, I called my health insurance and I was glad there was a human being on the other end. What we mustn’t forget with that — and I think this will also be a really important impact — is our succeeding generations. We always start from ourselves. What’s self-evident for us today, or seems strange to us, or where we feel, well, I’m glad there’s a human being on the other end — that’s going to look completely different for the generations growing up with this, who, in inverted commas, don’t know anything else. And we’re already seeing that with all these social media developments and so on. I’m not saying that’s good — that would again be a value judgment, and if we start analyzing and reflecting on that, we can do all of that, we can discuss it. Nevertheless, that doesn’t change the fact that, whenever a new generation grows up with new tools, techniques, and so on, they have completely different views than what we grew up with. And what we might struggle with, a succeeding generation might not struggle with at all, because for them it’s completely natural. As I said, you can question that, you can look at what that means for society — that’s a different level of discussion, and we can think about what we want and what we don’t want. And that’s entirely legitimate and sensible to discuss too. But we shouldn’t automatically assume that the way we see and judge things today is how succeeding generations will see and judge them too. They’re going to experience some of this quite differently. And with that, over to Sabine.
Yes, a fourth point occurred to me — Thursday was a public holiday and I was lying on the terrace in the sun, reading “Atlas der Entscheider” [“Atlas of Decision-Makers”], and there was an AI topic in there that I found presented so simply. There was just a flowchart, a kind of bundling of which task goes where, and it was simply, okay, if it’s 100 percent accurate, no errors, it goes to AI anyway. If it’s, let’s not worry about the exact percentage, but if it’s, say, 10 percent human involvement needed, then it takes a detour and gets checked again — how important is that 10 percent. If, in the end, it’s not that important, it still goes to AI. And then there are the tasks where you say, no, that doesn’t go to AI. And I think that’s simply a picture of how it’s going to work in companies in the future. That’s the basis for management consulting — when consultants go into a company today and say, okay, what can go in the direction of AI here. And that’s reality — I don’t think we even need to keep pondering that, that’s just reality.
What else did I want to say — oh right, the name came back to me, Snapchat, but it doesn’t matter what the tool is called. I think we’re also being asked as parents here. I’m firmly convinced that with my oldest — and my middle one too, actually — they already work with ChatGPT, I think they already have some sense about it, they’re already at that stage. With my youngest I’m not so sure, so you just have to stay in conversation with them and ask, what are you doing there, how does that work. I think that’s great though — if someone says, don’t go along with that, you’ve already understood that. And of course that raises the question, how close is the bond between parents and children? Where can we, because let’s be honest, our kids are far ahead of us in many ways — and how am I supposed to talk to someone about it if I don’t have a clue myself? That’s what I wanted to say. And then, of course, the topic of education — there’s hardly a day now where you meet friends and don’t talk about education. But why isn’t anything happening? I’m completely with you there, Layla, and it needs to go in, we don’t need to keep hammering knowledge into people anymore, we need to apply it. Except there too, I think, Layla, you’re probably a shining exception — a lot of kids are further along on this topic than their teachers. But why isn’t anything being tackled from the top? The last question I still have, listening to everyone earlier, is — shouldn’t AI be the solution to our skilled-labor shortage? Of course they’re not all at that level yet today, but I can make them a bit smarter, and then I’d have workers, wouldn’t I? That was a provocative question to end on. I wish you all a wonderful room.
Yes, I don’t think I need to say a closing word, because I thought the room was great overall, I thought the flow was great — that’s my closing word. I wish you all a wonderful weekend.
Yes, wonderful, thank you, thank you so much. We’ll just let that question hang in the air for now, to keep working on us, and with that we’ll move into the closing round. I’ve noted that, dear Sabine, and I’m happy to take that as a closing word too, and I’d actually like to take a look now at what your highlights from the room were, so to speak, in order to round things off and announce tomorrow’s topic accordingly. Marc, Jürgen, Layla, in that order please.
I’d like to be a bit polarizing again — our brain also has the morality of its “programmer,” among other things our belief systems and our environment, and the solution for me, for all these things, is always to question them. If that gets questioned, whether by AI or by us, then we’re already a good deal further along. And the last point I haven’t mentioned yet — AI is also incredibly expensive. ChatGPT currently has operating costs of 700,000 US dollars. So we also need to be clear that using this AI in the future is going to cost money too, and that’s a whole other topic — can every company afford that, and so on, and so on. It’s not going to stay free, it’s going to cost money. Thank you, thank you so much.
Jürgen. Yes, thank you all as always for your contributions. I’d like to somewhat violate my usual restraint to thank you specially once more, dear Yasemin, for your last remarks with the link to the scenario approach, using scenarios as a problem-solving cycle — and I think the topic of AI is predestined for the constructive application of exactly this scenario model, because it’s about specifying and shaping mental constructs for our future — what does the current scenario look like, the as-is scenario, already under the influence of AI today, as opposed to the expected scenario, in order to then be able to conclude, from that, the desired scenario, and beyond that, on a larger meta-level, eventually derived from the advantages and disadvantages we’ve derived today from the as-is scenario, the utopian scenario. In that sense I found this connection with the scenario model absolutely fitting. Thank you very much, have a nice Saturday, and maybe see you tomorrow, that’s still open for me.
Yes, wonderful. Thank you, thank you so much, Layla.
Yes, well, I have a lot of question marks, but fewer exclamation marks for it, and I find that quite nice already. I notice that we sometimes think we should, must, want to predict the future like a fortune teller, and maybe we simply can’t. But that’s okay too. And to close, I’d like to quote Bertolt Brecht from “The Good Person of Szechwan,” because I think it fits really well right now: “We stand ourselves disappointed and dismayed, the curtain closed, all questions still open.” I’m looking forward to more rooms. Thanks for listening.
Yes, wonderful, thank you, thank you so much. And to close, I’d like to bring things together and hold on to one thing in particular. First of all, I think it’s wonderful, Layla, when there are lots of question marks — I keep saying this — when there are question marks, we’ve started musing, started thinking, and that means we’ve opened up the possibility to sort our thoughts here, to go into reflection, to let it keep working on us, and maybe to discover new answers, new options, to find new answers for ourselves, or also to reaffirm old answers. That, too, is certainly a possibility, an option that’s valuable and important. And I’d like to briefly tie back into this point, because it’s been said several times today — we could basically make AI alone into a huge overarching theme. Yes, you’re right, maybe that’ll happen at some point too, we’ll see how it develops. My intention under this overarching theme, Pioneering Spirit with Brain, is actually to show the connections between things now. And you’ll see over the coming days and weeks — basically, we might be driven, to some extent, by the media too, our focus kept getting steered again and again toward something specific. But ultimately it’s about understanding the connections. It’s not AI on its own, or blockchain on its own, or VR, or whatever, considered individually, that’s going to bring about the change here at this magnitude — it’s the interplay of these technologies, which have already been running in the background for many years. Today we’re seeing the first results of that, which are now becoming noticeable to us too. And we’ve gotten a sense, through their application, of the direction this could go, or of everything it’s still going to be capable of in the future. And that’s coming too, much faster than we can sometimes imagine. And that’s exactly why it’s so important to look not just at AI on its own, but to consider the technological developments taking place together, and to think in combination, in order to realize what’s going to change, and at the same time to discover the potential for us in that.
And that’s something I’d also like to underline again to close, because that’s exactly what we started with this morning — namely, we tend to, because our brain naturally thinks first about securing our existence, what does that mean for us, do we even want that, and so on. And those questions are legitimate too, and they have their place, and they’re allowed to have their place. At the same time, we often think in either-or terms — is it important now to keep consciously steering our focus toward, what does this mean for me? What does AI mean for me? Where can I actually use it, apply it, unfold potential with it today already, and so on? And you’re welcome to ask yourselves that same question in the coming days when we talk about blockchain, about digitalization, about whatever else, without leaving out the societal or the macroeconomic dimension — we can gladly discuss that too. But we should discipline ourselves not to look too exclusively from one perspective or through one lens, namely the lens of worry, but also through the lens of potential. And if I also consider the question, what can I use this for, how can I use it for myself, then I can already do some really tremendous things with it today, things that multiply my potential, a thousandfold, at this point. Whether that’s AI, or other tools, or a combined version of these tools, is for now beside the point — everyone has to check that for themselves. And there are plenty of fields of application. Today I’ve brought a link again where you can use it, for example, to write a book — you don’t have to, you can, it’s an individual decision for everyone. I can only recommend that you really engage intensively with these topics, because, and this is what I said at the very beginning this morning, we tend to think linearly — that’s what our brain can do. We have difficulty thinking in exponential terms. But what’s happening now, through all these technological developments taken individually and then viewed in combination, is really an acceleration, an exponential increase to the power of, I don’t even know how much. We can’t even imagine that, our brain isn’t capable of that. It’s a bit like the compound interest effect — we keep hearing about it, sounds great, but basically our brain just doesn’t grasp it. No, the calculation just stops working at some point for us, we think linearly there too. And here we have an exponential increase happening right now. I don’t have the option of leaning back and saying, I’ll watch this, or I’ll deal with it in a year. I really need to look now and recognize what possibilities there are, how I can use this for myself, in order to get into this exponential growth, so as not to let the gap grow too large — and not as a negative scenario, but really, it’s going to be hard for every one of us to catch up here, because learning takes time. And if I let this time pass me by now — one person maybe three months, another six months, another a year — don’t forget, there are people on this market who have already been using and applying these technologies for years. And accordingly it’s all the more important that we keep asking ourselves, what’s the benefit for me? What can I use it for? How can I use it? And that’s my intention with these rooms, to really show these connections over the coming days and weeks — that’s why we keep reducing complexity to one topic area per room, so to speak, in order to then make the connections, to allow the corresponding discussions here, to open up the space for you to go into reflection, and, building on that, to make decisions. And not to postpone those decisions to “sometime,” but to get moving as quickly as possible and gather experience. I think that’s what it’s about — gathering experience, trying things out, making mistakes, and learning from those mistakes again. In that sense, thank you, thank you so much. We’ve run a bit over time today, please forgive me — but I think that’s simply because of the topic. I’m curious how things continue over the coming days. Tomorrow we first have the authors’ breakfast with Janik Gestör, and I’m really looking forward to that too, tomorrow at eight. You’re warmly invited to join. It’s about, “I’m So Free” — getting out of the hamster wheel and into the right job. And that will also be really exciting under the umbrella of Pioneering Spirit with Brain — what actually is the right job, what’s the job of the future, in inverted commas, and, and, and — all these questions will be the topic tomorrow. I’m really curious. You’re warmly invited, and I can assure you, it’s going to be exciting. So try to be there if you can make it timewise. Until then — ciao.
