YT YouTube2Text

YouTube2Text → Transcripts → KI-Experte warnt: „Was 2027 auf uns zukommt, unterschätzen alle!“ | Prof. Dr. Moring

KI-Experte warnt: „Was 2027 auf uns zukommt, unterschätzen alle!“ | Prof. Dr. Moring — Transcript

by Clasen Talk · 7,431 words · 1,124 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:27Dear viewers, humanity is currently
0:28undergoing the greatest technological
0:30transformation in its history.
0:32Artificial intelligence and robotics
0:34are spreading rapidly, especially in
0:36the USA and China. Where does Germany
0:39stand in this technological race? Today
0:41, I want to talk about this with one of
0:42our country's leading AI experts. He is
0:44a professor of digital economy,
0:46innovation and artificial intelligence
0:48at the International School of
0:49Management in Hamburg. Welcome, dear
0:51Professor Andreas Moring.
0:55>> Yes, thank you very much for being here
0:55. I'm glad I was here. Dear Professor
0:59Moring, let's get straight to the point
1:03. Where do we stand in this revolution,
1:06and what is it that actually lies ahead
1:09of us? How should we imagine it? What
1:13kind of dimensions are we actually
1:15talking about?
1:18>> Okay, that is quite a lot of questions
1:20all at once; in terms of the actual
1:22dimensions, it is comparable to the
1:23scale of the industrial revolution. So,
1:27we're not talking about a repetition of
1:29the kind we know from the millennium
1:32books with docom and digitalization,
1:34but it goes well beyond that, because
1:37digitalization and the Internet were an
1:39extension of what we knew. With
1:42artificial intelligence, we now possess
1:43entirely new machines, much in the same
1:45way that the steam engine was a brand
1:47new innovation back in the day. Where
1:49are we at the beginning of development?
1:54AI has been around for a long time, but
1:56the development we have now is that we
1:58are in the first quarter, maybe the
2:00first third, but definitely not further
2:02along. We are all
2:05>> familiar with those diagrams showing
2:07exponential growth; it begins quite
2:09gradually, then it rises slowly, until
2:11that famous curve is reached and it
2:13suddenly shoots straight upwards.
2:16>> Yes, exactly.
2:17>> Where would you classify us? So we are,
2:21this is what they are calling the
2:22hockey stick, isn't it? Well, we are
2:25basically at the bend, which is now
2:27leading up to this steep curve, pretty
2:29much at the beginning. And the years
2:332027 and 2028 are going to surprise us
2:35with all sorts of things. Um, compared
2:40to that, what we've experienced in the
2:42last two years will probably be-I don't
2:44want to say it sounds ridiculous, but
2:46it will be another giant leap forward,
2:48because now all the computing power is
2:51there, the models that support the
2:53whole thing are extremely sophisticated
2:55, but they are still the worst models
2:57we'll probably have, because, as you
2:59said, they'll just keep getting better
3:01and better and better. And these will
3:05be huge leaps in the next year and the
3:07year after that alone, which we
3:08probably can't even imagine. But the
3:11path is set.
3:13>> I mean, having the imagination is
3:16really a challenge. Um, when it came
3:20out, it was already an awakening moment
3:23for many, a moment of shock, um, it
3:26depends on the perspective. But what
3:30exactly will we experience in 2027 and
3:322028 that will finally give us a
3:34concrete idea of what is going to
3:36surprise us, so that we can better
3:38understand these events?
3:41>> For example, what you see behind you is
3:43robotics and AI, which is the next big
3:46development, which is no longer just a
3:48pipe dream taking place in some lab.
3:51Instead, you can see behind you that it
3:53is already a reality in China, for
3:55example, and that sounds like it at
3:57first glance. Well, robotics and AI,
3:59okay, we know robots, but they are
4:01robots that have learned to do
4:03everything that humans can do and much,
4:06much more, and above all, that continue
4:08to learn continuously. That is the big
4:11lever. So, whatever a robot learns
4:13somewhere in the world, every other
4:15robot will know about it just one
4:17second later. And then there will be a
4:21huge explosion in skills. Then there
4:25will also be further breakthroughs in
4:27AI, which we know from computers. In
4:30other words, AI models. We now often
4:33hear reports that AI has solved
4:38physical or mathematical problems that
4:41humans have struggled with for 90 or
4:43100 years. And they're getting there.
4:46We are going to see a great number of
4:47breakthroughs in the field of medicine.
4:50This is a huge area that is still very
4:53much underestimated. For example, new
4:56medications that cure diseases that we
4:58haven't made any progress on for
5:00decades or generations. And then there
5:04is certainly also the growing
5:05autonomization and automation within
5:07our various jobs. This primarily
5:10affects desk jobs, where an increasing
5:11number of people are being replaced by
5:13artificial intelligence. Sometimes
5:15completely, sometimes to a large extent
5:17. So that means: Everyone who, like us,
5:20sits here at the computer and mainly
5:22works there, will have to prepare for
5:24the fact that their jobs will
5:25definitely change significantly and
5:27that they will no longer do what they
5:29have done before, but that many things
5:31will be taken over by AI and that many
5:33new things will come their way, where
5:35they will have to relearn and relearn
5:37–not only how to deal with AI, but
5:39also how to perform completely
5:40different tasks that were not part of
5:42their job before.
5:44>> Hmm. And are you rather optimistic
5:49about this or do you sometimes get a
5:52feeling of helplessness or even fear?
5:56Because there are always news stories
5:58going through the news forest, for
6:01example, that people are afraid that
6:03humanity could even wipe itself out.
6:08And such catastrophic scenarios are
6:11sometimes portrayed. Is this a lot of
6:14PR? Is this an exaggeration by the
6:17media? Are people trying to make
6:19themselves important or is there a real
6:22background to it? Both.
6:26>> So what we've seen in the last few
6:28weeks was already PR, it was also a PR
6:30campaign. Open AI, Chen GPT, and also
6:33Tropic from Claud–that's the other
6:35big provider–they want to go public,
6:36and that requires a lot of attention.
6:40And well, how can you get more
6:42attention by briefly announcing that
6:44humanity is on the verge of extinction,
6:46to ultimately demonstrate the power of
6:49AI, or at least to pretend to, and to
6:51simply get yourself into the
6:53conversation. And that was indeed the
6:58point, to obtain regulation, not
7:00because AI is inherently so dangerous,
7:02but ultimately to serve as a form of
7:04protection against our Chinese
7:06competitors. Yes, we're talking about
7:08American companies. And this regulation
7:11would be of benefit to them and help to
7:13keep the Chinese out of the market. But
7:16there is a risk. There is also a
7:19distinct risk that humanity could be
7:20wiped out, although not because the AI
7:22suddenly goes crazy, but rather, as it
7:24always happens, when there are people
7:25who choose to use the AI for things
7:27that are capable of having truly
7:28catastrophic consequences for us all. I
7:31personally do not have any great fear
7:33regarding the power of AI, because it
7:36is ultimately nothing more than a model
7:38and a series of complex calculations.
7:43But as you say, in many areas it is a
7:45bit of an insult, right? When you
7:49realize that AI suddenly makes things
7:52faster, better, and much, much easier,
7:55things that we have to work hard for
7:58and think hard about in order to
8:01protect, then it can be a bit stupid
8:04sometimes, right? So you find yourself
8:07thinking, okay, what is my performance
8:09even worth now that this machine has
8:10suddenly appeared on the scene? But
8:12that is not something to be afraid of.
8:13Yeah, I'm not afraid of it, and nobody
8:15should be either, but it's normal to
8:17feel a little offended, right?
8:20>> Mm. Yeah, but have we already passed
8:23the point where artificial intelligence
8:26is smarter than the average person and
8:29that we're already experiencing in my
8:33finger 27 28 that there is such an
8:35artificial superintelligence that can
8:39achieve many times more than normal
8:41mortals? Yeah, it might be easier for
8:45me to achieve this goal than for you or
8:48others, but that also fundamentally
8:51raises the question of how to deal with
8:54a superintelligence that is, so to
8:57speak, smarter than humanity.
9:01>> Yeah, that's the big question. So, the
9:03first part of the question, has AI
9:05already overtaken us? Yeah, not
9:08everywhere, but in many areas. You
9:10always have to be careful there. That's
9:12the thing with the offendedness we had
9:13earlier. Of course, AI can calculate
9:15better. Yes, AI can effectively
9:17memorize things faster than humans can,
9:19and then it is able to replicate them,
9:21which is to say, it can imitate those
9:23things. AI is able to analyze massive
9:26amounts of data much faster than we
9:27humans, who would otherwise surrender
9:29before large Excel sheets and then no
9:31longer be able to make any sense of
9:33them. We spoke about AI combined with
9:35robotics earlier, and it has now also
9:37become capable of running a marathon
9:38faster than any human being. That was
9:39also in China. It was the first
9:41instance in which a robot completed a
9:42half marathon in a faster time than a
9:44human being could manage. The human
9:45wasn't even out of breath afterwards.
9:47Yes, people almost collapsed afterward.
9:50Yes, AI has already overtaken us in
9:52many areas. That will also be the case
9:55in other areas. Everything that is
9:58somehow regulated and based on a lot of
10:00data is where AI will eventually
10:02overtake us. The good news is that
10:03there are still many other things out
10:05there. AI can't overtake us because it
10:09is not in the real world, because it
10:10has no emotions, because it has no
10:12memory, and so on. And in that respect,
10:16there will still be enough remaining,
10:18but the superintelligence, if it does
10:20indeed exist, is now no longer that far
10:22away in those specific areas that we
10:24have been discussing during our
10:26conversation here. The question is, um,
10:29that's a bit of a philosophical
10:31question, will we get it? Well, the
10:34notion that many people have regarding
10:36the nature of superintelligence is one
10:37that is somewhat reminiscent of the
10:39Terminator films. Yes, at some point,
10:42um, the AI will come along and say, I'm
10:44going to subjugate humanity and you're
10:46my slaves now. That won't happen. If
10:49there were an intelligence that was
10:51truly smarter than all of humanity, my
10:54theory is that it would not be so
10:55foolish as to reveal that to us,
10:57because we might then find a way to
10:59pull the plug after all. Instead, such
11:02a form of intelligence would simply
11:04manipulate us in such a manner that we
11:05would not even be aware of it happening
11:07at all. Yeah, really exciting, because
11:11when you're in the zoo and you see
11:13chimpanzees, that's only a tiny
11:14fraction that's different from us. Then
11:20we end up looking at them very
11:23condescendingly. And when I see the
11:26guys behind me, the robots, who will
11:29eventually be equipped with super
11:32intelligence, I'm really curious to see
11:35how they look at us and how they
11:37perceive us. Is there actually any
11:43research or ideas about how they
11:45perceive us? Yes. There are, let's say,
11:51various speculations and models which,
11:54however, also lie more within the,
11:56shall we say, physical domain or
11:59perhaps the philosophy of physics, or
12:01even in the realm of computer
12:03philosophy. Nobody knows, because we
12:08don't really know how a dog perceives
12:10us, or a chimpanzee. But I believe that
12:14they will probably find us to be quite
12:15interesting, primarily because we are
12:16the ones who have been responsible for
12:18creating them. So I think they will
12:20then ask themselves, who was it that
12:22actually gave us life, so to speak? Um,
12:25I think they will be very interested,
12:27um, and then maybe lose interest at
12:29some point and say, okay, do we get it?
12:32We are now heading out into the
12:34vastness of space, doing that which
12:36humans are unable to do, which is to
12:38say, essentially flying off somewhere
12:40ourselves.
12:41>> Yes, that sounds totally crazy, as if
12:43these machines are becoming autonomous
12:45or coming to life. But if you look at
12:47the situation in China now, we are
12:49already seeing the rollout of
12:51autonomous driving systems. Um, we have
12:55dark factories, where robots are
12:56already starting to produce robots
12:58themselves. Um, and that really seems
13:02to be just the beginning. I recently
13:05saw a video of a fully automated food
13:09kitchen, um, where the food is prepared
13:14and delivered by robots. Is that, so to
13:18speak, the future? What we saw in those
13:21prototypes in China, is that also
13:23something that we will encounter in our
13:25everyday lives in Germany at some point
13:28in the future? Yes, so it's probably
13:30not the future, but exactly as you just
13:32described, already the present. Just
13:34not in Germany yet. In the USA and in
13:37China we are already there, or rather,
13:39they are already two years ahead of us,
13:42and naturally this will also prevail,
13:44just as other technological innovations
13:46, such as the smartphone or the
13:47internet and so forth, have also
13:49managed to establish themselves in our
13:51society over time. It was first in the
13:54USA and now it's worldwide, and here
13:56it's the same and that's exactly what
13:57you're saying. As we already mentioned
14:01earlier, AI is now coming out of the
14:03computer and into the real world, and
14:06that affects all areas, professions,
14:08and activities that take place at some
14:11desk, and subsequently, others as well;
14:13although, one must note that these
14:15tasks are those which can somehow be
14:18standardized, wouldn't you agree? As
14:22you mentioned regarding the factory,
14:23there are essentially standards and
14:25clear distances, and it is all
14:26assembled in the kitchen, isn't it? If
14:29everything is where it should be, a
14:31robot can do it just fine. Hotels, for
14:35example, cleaning rooms, folding beds,
14:38and so on. That's all standardized.
14:43Maybe it's also used on construction
14:45sites. Yes, it also depends on what
14:46you're building and how you're building
14:48. On other construction sites, it's not
14:50. Yes, because those are individual
14:51houses, custom-made items, which means
14:53that the components are not
14:53manufactured according to standard
14:55norms. But everything where there's a
14:57standard, where there's a norm, and so
14:59on, there's a lot of that in Europe and
15:01Germany. Of course, that's a great
15:04field of application for this robotics
15:07AI, which may not be part of everyday
15:10life here in Germany in the foreseeable
15:13future, but will definitely be used in
15:15industry and services.
15:18>> Yes. Um, and if you compare Germany
15:23with the USA and China, perhaps Europe
15:28as a whole, where do we stand?
15:34>> We are at the back of the line compared
15:36to China, compared to the USA, and also
15:38compared to Israel, for example, who
15:40are also very far ahead in that regard.
15:45We also observe this in other nations
15:47or smaller countries, taking Ukraine as
15:49an example; even if the circumstances
15:51are certainly not pleasant,
15:54>> we are currently learning from their
15:55experiences regarding how to
15:56effectively utilize drones in
15:56operations. Yes, and on a much larger
15:58scale, this is also the case with China
16:00and the USA. Why? Yes, because they
16:04have powerful AI models. We are
16:06familiar with Chat GPT, and perhaps
16:08some of you are also familiar with
16:10Clord as another model. However, there
16:13are equally powerful models in China,
16:16like Deep Seek and others, which are
16:18all called AUUR, meaning that everyone
16:20is able to make use of them. Yes,
16:23whether you want a Chinese model, that
16:25is, everyone has to decide for
16:26themselves. Um, but that means they are
16:30basically given away for free, yes, and
16:32anyone can take them, and of course
16:35that accelerates the progress and the
16:37gap and the lead even further. That
16:41means that with the current technology,
16:45we as Europe have no chance of catching
16:47up. Um, is that the only chance, or are
16:49there still several other chances
16:50available? Um, one chance would be that
16:52at some point there will be a
16:54technological breakthrough, yes,
16:57>> that AI works completely differently
16:59than it does today with these language
17:02models, um, and niches, yes, niches
17:04where we are good, because these large
17:07models are such general models and as
17:09always, yes, if you find a niche in the
17:12market, then you are not the global
17:14leader, but you are good in your niche
17:17and, um, that could be an opportunity.
17:22But that massive big business is
17:24essentially controlled by the Americans
17:26and the Chinese, if I am understanding
17:28them correctly, that is to say. And we
17:30have no chance at all, well probably
17:32not, unless there is a major
17:34technological leap that allows us to
17:36catch up once again. How come, how do
17:40you explain that? Because ultimately,
17:43all the major industrial revolutions
17:47did originate in Europe, such as the
17:49steam engine and others, but now the
17:50old continent seems to have been left
17:52behind in the current global race.
17:56>> Well, one thing is certainly not the
17:58cause, and that is a lack of
18:00inventiveness or research and
18:02development potential, because many
18:06things have indeed been invented right
18:08here in Europe. But, as you say, they
18:10weren't implemented here. Well, what do
18:12you need for implementation? You
18:15require a great deal of capital, you
18:17require a vast amount of data, you
18:18require a significant degree of freedom
18:20, and you need support; furthermore,
18:22you require cooperation with the
18:24economy, which is to say, collaboration
18:25with established corporate enterprises.
18:28That works very well in the US. There's
18:31a lot of money there. It's the same in
18:33China. There is also a significant
18:35amount of capital coming from private
18:37companies, but also from the state,
18:38which is investing quite heavily
18:40>> in China.
18:41>> In China. Yes, starting from the
18:43research phase all the way to the
18:45implementation and so forth. You need
18:48>> gigantic sums of money. I want to jump
18:51in right here, because we are not
18:53talking about 10 billion euros here;
18:56>> rather, we are talking about amounts in
18:58the trillions when you add everything
18:59together.
19:00>> Yes. Tesla, led by Elon Musk, is
19:05currently investing 120 billion euros
19:08in a chip factory in Texas. That sum
19:11exceeds the entire German investment
19:14despite all increases,
19:16>> and this facility will eventually
19:17become the largest chip factory in the
19:18world by a significant margin. But that
19:20also shows what challenges we still
19:22face, especially with regard to
19:25computing capacity, computing storage
19:27capacity, and energy. Energy also seems
19:32to be a key issue as an infrastructure
19:35to advance this digital ecosystem.
19:40>> Yes, exactly. Those are the points that
19:43come up. So, money and investment and
19:46infrastructure, the money in the US is
19:48there. Yes. It is always being said
19:51that this is just a dot-com bubble once
19:52again revived for us today. No, no,
19:54there is really investment in
19:56infrastructure, exactly as you said,
19:58and in large quantities. Also, data
20:00centers are as big as Manhattan. Yes. A
20:04data center, or exactly the same
20:05dimension you mentioned, is also being
20:07operated in the very same way over
20:08there in China. Huge data centers are
20:11being constructed and, just as you have
20:13pointed out, we need energy—
20:14specifically, affordable and reliable
20:16energy. And that's where we have a real
20:19problem in Europe and especially in
20:20Germany. Every one of our viewers knows
20:23just by taking a glance at their
20:24current electricity bill exactly how
20:26expensive it has become for all of us.
20:28Yes. And it has to be worth it. So if
20:30the federal government is now claiming
20:32that they are launching a data center
20:34offensive, well, I would be quite
20:35curious to see how that offensive
20:37actually functions, especially
20:38regarding building permits and
20:40everything else that is involved in the
20:42process. But the most beautiful,
20:43largest data center is of no use to me
20:45if the electricity is so expensive that
20:47I can't afford it. And that's a huge
20:50problem. And then there are also the
20:52various regulations, keyword protection
20:54measures, and so on to consider. Yes,
20:57the EU AI Act is always praised by the
20:59EU, yes, exemplary, exemplary, but it
21:01is a major obstacle, yes, because you
21:04have to consider so many things and are
21:06not allowed to do them. And this
21:09combination of not having enough money,
21:12no available infrastructure, the lack
21:14of truly affordable energy that we need
21:17in large quantities, and, on top of
21:19that, a tight corset of regulations
21:21that we are constrained by. You can't
21:26really make any progress with that,
21:27right? It's like a 100-meter leek with
21:29a few weights.
21:33>> But are we not actually risking our own
21:35viability for the future here in Europe
21:36by doing that? These are groundbreaking
21:40changes, and if we cannot participate,
21:43we will face enormous economic
21:44competitive disadvantages in the long
21:47term, especially if the so-called dark
21:49factory engages in mass production and
21:51we are simply unable to keep pace with
21:53those developments and stay relevant in
21:55the current market environment. I
21:58believe that this is going to have a
22:00negative impact in many different areas
22:01as well. Yes, we can already see that
22:06the German economy is not doing
22:07particularly well, which has a lot to
22:09do with energy prices, as well as
22:10regulation and other various factors of
22:11that nature.
22:13>> And then there is the fact that twenty,
22:16thirty, or forty years ago, that would
22:18have been, shall we say, rather stupid
22:20and not all that pleasant either. But
22:25there was still time to catch up, and
22:27as we noted at the very beginning, the
22:29hockey stick curve—the way that
22:31development is accelerating—that is
22:33the time we simply do not have at our
22:35disposal any longer. So, this change is
22:39not happening over 20 decades, but
22:40rather over a period of 5 years. And if
22:44you fail to react within those 5 years,
22:46then you are, for the most part, simply
22:48out of the game entirely. Unfortunately
22:51, that has to be said. Many people have
22:53not yet understood that. In business,
22:56yes, but in politics, I don't get the
22:59feeling that there is an awareness of
23:02how tight time is.
23:05>> What should politicians do? First of
23:10all, the restrictions that have been
23:12imposed by the regulation are now being
23:14lifted. That would be the easiest thing
23:16to do. One can very quickly repeal or
23:19even suspend laws and then simply
23:21invest or have others invest in that
23:24which we were discussing earlier
23:26regarding the infrastructure, yes, and
23:29also the development. Another idea
23:32could be, let us say, to formally
23:34declare the creation of special
23:35economic zones. It does not necessarily
23:38have to be limited to Germany, it can
23:39just as well be located within Europe.
23:42To say, okay guys, listen, all these
23:44regulations don't apply here. And you
23:48now have a free hand for 10 or 5 years,
23:52right? Go ahead and develop it. You are
23:58now free to simply do exactly what is
24:00necessary, and you can be assured that
24:02the political sphere will not put any
24:03obstacles or difficulties in your way
24:05as you proceed. Yes, politics doesn't
24:09really have to do much in terms of
24:11funding or anything like that, but
24:13simply remove the hurdles that we
24:15listed earlier, and many of them are
24:17also possible, right? Other options
24:21could also be possible, but then it
24:22might become necessary for one to
24:24occasionally call certain convictions
24:26into question as well. Yes, keyword
24:28affordable energy. If we are aware that
24:32this is what truly matters, then one
24:33really has to consider whether the
24:34entirety of this energy transition
24:36policy actually makes any sense in its
24:37current form. Yes, but it just doesn't
24:40fit together.
24:44>> Well, a reliable energy policy is of
24:46course a basic requirement and then at
24:49globally competitive prices, but we are
24:52still a long way from expanding the
24:54grid. This also applies to the hubs in
24:58Frankfurt. I was also surprised that we
25:03can hardly manage to add additional
25:05connections there.
25:07>> Hmm. But do you have an explanation?
25:11You were talking about politics. We
25:15have had digital ministers and so on
25:16for many, many years now, and there
25:17have always been repeated appeals as
25:19well. Uh, what's the problem here? Why
25:25can't we get the horsepower onto the
25:27streets?
25:30>> Well, it's largely a matter of
25:32understanding. Uh, not because it's
25:35somehow so difficult. Mm. but because I
25:38believe that the focus, when it comes
25:40to many of these political issues,
25:42actually lies somewhere else entirely.
25:44So AI is still perceived as, well, just
25:46some sort of new technology that has
25:47been integrated and is now somehow
25:49existing somewhere inside of that
25:51system. Yes. Uh, but otherwise it's not
25:54really relevant. The extent of it is
25:57simply not known to many, or indeed to
25:59the vast majority of people, as is the
26:01case. Uh, because, as I said, the focus
26:04lies somewhere else. We see it in the
26:07discussions, let us consider what is
26:09being debated, yes, about Ms. Bars and
26:11so on and so forth and who said what
26:13here, and for how much longer are we
26:14going to struggle with the details of
26:16some kind of pension reform that has
26:18been proposed? Yes. Uh, and so on. Well
26:22, I think that's just the system, and
26:27also in terms of perception, right?
26:32Human beings are wonderful when the
26:33Chancellor travels to China, just as we
26:34see him in the picture in the
26:35background while he is looking at the
26:37robots. But I think the leap, okay,
26:40what does that actually mean for
26:41Germany? It's somehow not really there
26:44yet. Yes. In that respect, as I have
26:47mentioned, I am not feeling very
26:49optimistic regarding the political
26:51situation. And there's another thing,
26:54and this is not only the case in
26:55Germany, but also in the US, where we
26:57see a certain front from the left
27:06forming against AI. There are more and
27:09more demonstrations against these data
27:10centers that you mentioned. There are
27:12an increasing number of demonstrations
27:15against the idea that AI is taking away
27:17people's jobs, in quotation marks. So,
27:20this is a new leftist narrative of the
27:22new class struggle, which is, yes, a
27:245.0 version of it, so to speak. There's
27:28capital and these are the evil, um,
27:30technology people and they're taking
27:32your jobs away, they're making energy
27:34more expensive for you, um, and they
27:36want to control and subjugate you all
27:38somehow, so to speak. And that's why we
27:40have to prevent that now and that's why
27:42no data centers are being built here
27:44and everything is being blocked and so
27:46on. So, so to speak, that short-term
27:50thinking is triumphing, right? For
27:53there is no example in the history of
27:55all of humanity where technology could
27:56have ever been successfully stopped or
27:58held back. So, you can't stop it. The
28:00question is, how do I deal with it? And
28:03being against it and trying to stop it
28:05somehow really gets to the bottom of it
28:08in the long run.
28:10>> Yes, but here in Germany, we are
28:12discussing taxes on sugar or plastic,
28:14or perhaps the 85th reform, which is
28:16intended to be particularly extensive
28:18for pensions or elderly care. Yes, but
28:21the big picture, we do not seem to have
28:23fully grasped that yet, the fact that a
28:26revolution is currently underway which
28:29is destined to change the world in a
28:31truly radical way for all of us. And of
28:34course that causes fear. You also
28:37mentioned that, regarding the
28:38resistance that is seen in parts of the
28:40United States. Of course, we don't know
28:44how it will end, but I don't think it
28:50will have such a broad impact. In any
28:52case, I don't have that kind of impetus
28:54, because the construction work will
28:56continue. And the only question is,
29:00when do we start getting involved? In
29:03other words, to really recognize the
29:05opportunities that this offers. On the
29:09other hand, one could certainly argue
29:10that, yes, certain jobs are at risk,
29:12but one could also view that situation
29:13as an advantage as well. We recently
29:18had a top manager from the copper
29:19industry here, they have blast furnaces
29:21. If robots take on the dangerous, hard
29:24work in the future, then that should,
29:25all things considered, be an advantage
29:27for us. And you can apply that to many
29:32areas of old care systems, for example.
29:37The Chinese are also heavily betting on
29:43finding very specific applications that
29:46will ultimately make people's lives
29:50easier. Or is that being too optimistic
29:54?
29:54>> No, it's a question of perspective. As
29:56always with technology, right? There
29:58are negative effects, so to speak, and
30:00naturally there is also a significant
30:02amount of positive potential to be
30:04found as well. Regarding what you just
30:06said about dangerous jobs, yes, it is
30:08in any case a huge advantage if people
30:10no longer have to perform them, and if
30:11there are no more accidents occurring
30:13as a result, and so on and so forth
30:15throughout the entire process. Then, as
30:17we also mentioned earlier, the advances
30:20that we will soon see in medicine, that
30:23people will no longer have to die from
30:26diseases, as has been the case up to
30:28now. Of course, I don't know where
30:31there will be breakthroughs, right? In
30:34cancer research, yes, perhaps also with
30:37AIDS, and other diseases. Then of
30:41course also in the whole area of
30:44sustainability, climate protection and
30:46so on. AI can make huge contributions
30:51there. Yes, to understand how it works?
30:54But of course, this also applies to the
30:57development of new fuels, new materials
30:59, and insulation, all the way to the
31:02optimal management of traffic
31:03activities in order to reduce the CO2
31:06footprint, or the implementation of
31:08AI-controlled systems in buildings to
31:10ensure that no further energy or heat
31:13is wasted, and so on. And of course
31:17also in research, yes, in education,
31:19you have to look at where you can use
31:22it now, yes, and how. You can also go
31:27through all the natural sciences, where
31:30we will have new insights and from that
31:33in turn new, new inventions will result
31:35, yes, which of course make life much
31:38more convenient and much easier and
31:40which in turn offer potential for
31:45marketing. Yeah, so if I have a good
31:48idea, a good product, a good service,
31:50then that's actually already a business
31:52model. And there will be demand for
31:56that. Then there's another question,
31:59whether there will still be enough
32:00people who can afford it. Uh, no, who,
32:03so to speak, earns the money here and
32:05how. So, if we now have this dystopian
32:08idea that virtually everyone earns
32:10their job because only AI is working,
32:12then of course it will be a bit
32:13difficult. Uh, but Elon Musk is
32:16pursuing this theory and is therefore
32:18calling for a kind of basic income for
32:20everyone.
32:21>> Hmm. Yeah, that's the idea, right? And
32:25he even says a universal high income,
32:27so that everyone gets a lot. Okay, okay
32:31. Uh, Mask likes to come out with
32:32something. Uh, in that respect, it's
32:35certainly also a bit of PRA, which we
32:38also had at the beginning.
32:40>> Yes, that's of
32:41>> course the question. I don't think that
32:44we're going to lose all our jobs
32:46because AI is taking over everything.
32:50But still, and here we're back to the
32:52political question, it's of course a
32:55huge challenge. If, for example, the
32:58unemployment rate were to double
33:00compared to current levels due to the
33:02impact of AI, because certain types of
33:04jobs are simply no longer being
33:06performed, then we would face a most
33:08massive problem regarding our social
33:10security system, our pension system,
33:12and everything else associated with it.
33:15It's already not stretched to the limit
33:17, but is already being financed with
33:19debt. And in that respect, we don't
33:24need mass unemployment here, but a few
33:28million more in unemployment due to AI
33:32will definitely cause it to collapse.
33:38>> And we're talking about a period of the
33:40next 10 years, right?
33:42>> Yes, that's
33:44>> realistic, right? Yes, definitely
33:47>> 2035,
33:48>> and companies are gradually starting to
33:54stop hiring people or filling vacant
33:56positions and optimizing processes.
34:00>> You can see that in many areas, and of
34:03course it has its disadvantages, but
34:06isn't there also the possibility that
34:09new jobs will be created as a result?
34:15When do you think the Chinese will roll
34:18out armies of robots from their
34:20container ships to help us with
34:23household chores, for example? And
34:27someone has to sell them, someone has
34:29to maintain them. There should actually
34:32be new business models behind them.
34:36>> Yes, I expect that to happen around
34:392030. Because they will first serve
34:42their domestic market. And when that is
34:46saturated, they will then start
34:48exporting to Europe. That's exactly the
34:50same thing that is happening with cars
34:51right now. First, they sold all their
34:54electric cars in China. Now everyone
34:56has one and now they're bringing it to
34:58Europe at dumping prices. And then VW,
35:00BMW and so on have a real problem. And
35:03it will be the same with humanoid
35:05robots. It won't happen tomorrow, but
35:09it will happen around the turn of the
35:11decade with the jobs created by AI. In
35:15any case, we've always seen that with
35:16technological innovations, that
35:18completely new jobs are created that
35:19didn't exist before. Others will
35:22naturally disappear. And it will, we
35:27already had this in the beginning, it
35:29will simply lead to people needing
35:32different skills and abilities in their
35:34jobs. So, the skill or competence, like
35:40creating a PowerPoint presentation or
35:42knowing Excel formulas and messing
35:45around with them. Well, that's no
35:48longer a skill, right? So, AI makes
35:50that much faster. Also, what used to be
35:54really highly paid, yes, good design or
35:58controlling, reporting and so on. Also
36:03middle management, tracking certain key
36:06figures and so on, optimizing processes
36:08. Uh, yes, AI can do that much better.
36:14Uh, in general, and then also agentic
36:18AI. So agentic AI is basically robotics
36:21, only on the network, in the computer,
36:23which also acts completely autonomously
36:25. It can place orders, it can create
36:30annual financial statements, it can do
36:32quality checks, it can conduct customer
36:34meetings and so on and so forth.
36:37>> I find that a really interesting point.
36:41Uh, uh, AI agents. How should we
36:44imagine that? So, as a private
36:46individual, I will then have five
36:51agents in the future, who are there to
36:54do my tax return, manage my bank
36:57account and invest a sum X on the stock
37:04exchange. The next one takes care of
37:07the children's homework and the third
37:11one thinks of how to surprise his wife
37:14for their wedding anniversary with a
37:17trip to Honolulu, I don't know.
37:21>> Exactly.
37:22>> Yes.
37:24>> Or the Black Forest. Yes, exactly. So
37:26basically yes, in the examples you gave
37:29, I would say that some of them I would
37:31definitely hand over to agents, others
37:34not. So I would leave creativity to
37:37people, including supporting children
37:40in learning. So there are some great
37:42applications. But I wouldn't leave it
37:44completely to the AI, but as you said,
37:46it's right. I myself have, I think,
37:50more than 20 agents who do things for
37:53me. Yes. Yes. I built them myself. It's
37:58much easier these days. You don't even
38:01have to be able to program them anymore
38:02. The AI has come so far that they can
38:05do it with Chat GPT in the work mode.
38:08Um, you can now switch between Chat,
38:10Work, or Codex at the top, but that's
38:13already included in Chat GPT. If you
38:16click on it at the top, you can click
38:18on Chat GPT and then Codex again. Um,
38:22and then you just have to tell the AI,
38:23in quotation marks, what you'd like to
38:25have. Um, and then it builds this
38:28automation or agent for you. It's not
38:30quite that simple. Um, because we need
38:33to know exactly how the technology
38:35works and what we actually want. Yes,
38:38we need to be able to describe it very
38:39precisely and in detail. But we don't
38:41need to be able to program anymore. And
38:43then I can build agents for all sorts
38:45of things, for the stories you've told,
38:47or for other things. And they do it
38:51either when I tell them: "Okay, go
38:54ahead!", or at a specific time. And
38:57then they just do their jobs, their
38:59tasks. And if they can do that for us
39:03privately, then you can do that in your
39:05professional life and in your everyday
39:07work, and of course that's simply much
39:09cheaper than paying people for it,
39:11right? People also have other skills
39:15that AI doesn't have, which then
39:17perhaps, or most likely, my theory,
39:19will also become more valuable in terms
39:21of perception, in terms of appreciation
39:23, but also more valuable in terms of
39:25the question of what people are
39:27actually paid for? That's a really
39:29exciting question. What kind of skills
39:32are these and where can they be used?
39:36>> Yes, so what skills are these? Maybe
39:41this year, maybe early next year, I'll
39:43release a new book that deals with
39:44exactly that. And I've narrowed it down
39:48to ten skills, because there are a lot
39:51of studies that list 30 skills and so
39:53on, but no one can describe them. So,
39:57it's data literacy and AI literacy, so
39:59to speak, application skills with data
40:02and AI, right? Can I do this, can I
40:05train, can I take courses and so on.
40:07How do I use it correctly? So data
40:09literacy, AI literacy, then it's
40:11systemic thinking. Um, AI can't do that
40:15. Yes. Um, but we humans can do it very
40:17well, because we are also biological
40:19beings and live in systems and cycles.
40:22No one is a biological being, doesn't
40:24live in cycles and so on. So,
40:26>> can you explain what systemic thinking
40:28is a bit more specifically, can you
40:29give us an example? Um, systemic
40:32thinking simply means understanding,
40:34for example, how the annual cycle works
40:36. So why, um, spring, summer, autumn
40:39and winter and how all of this is
40:40connected. Yes, um, that the snowflakes
40:43fall from the trees, that the animals
40:46hibernate and that, um, that the water
40:49can freeze and so on. Um, these are all
40:52, um, and that it's because we have a
40:54different position to the sun and so on
40:56, right? Um, well, there are also a lot
41:01of, let's say, irrationalities in these
41:03cycles, which are not statistical, but
41:06have to do with other laws. Or another
41:10system is the family, yes, um, there
41:12you like some more, others less, and
41:14sometimes everything is fine and
41:16sometimes there are arguments based on
41:19trivialities. We humans understand that
41:22, right? We can't understand the
41:25dynamics in such a system, because it
41:27has to do with emotions, also with
41:29sympathy, with history and so on. AI
41:33just can't do all that. But that is
41:35extremely important for understanding
41:37companies, for understanding markets,
41:38for understanding customers and so on.
41:41So that's systemic thinking. Um, then
41:43we had creativity, right? Um,
41:47creativity doesn't just mean I have
41:49lots of ideas, creativity means I have
41:51ideas to solve a problem in a new way
41:53and better than before, right? Um, and
41:57K can't do that either, because I have
41:59to be able to put myself in other
42:00people's shoes and really understand
42:02the problem. Um, and also this famous
42:04out-of-the-box thinking, KIE is in the
42:06box. We humans can think out-of-the-box
42:08. Um, it's, um, the desire to learn,
42:12yes, curiosity, being curious and, um,
42:14wanting to learn. Um, it's a human
42:17ability that AI doesn't have either. It
42:20gets an order, um, and then it does
42:22that, but it's not curious, yes, um,
42:24and learns new things of its own accord
42:29. Um, it's, um, resilience, yes, um, um
42:32, for humans, being resilient means
42:35being able to stay focused, yes, um,
42:37being able to set goals, being able to
42:39stay focused. Um, it's also, um,
42:43self-motivation, used to be called
42:45discipline. Yeah, not so much anymore,
42:48today, there are other words for it.
42:51But ultimately, it's discipline, right?
42:53Discipline. Um, and then we also have,
42:57um, empathy. AI can't, right? Put
43:01yourself in other people's shoes,
43:03because I can't feel and take on their
43:05perspective. Um, it's also about
43:08communication and motivation, team
43:11motivation, reaching people through
43:14communication and motivating them and
43:17so on, giving them meaning. And number
43:2210, experiential knowledge and
43:24intuition, right? Um, some people like
43:26the term intuition, others prefer
43:28experiential knowledge. They're both
43:30the same thing. So, um, what we don't
43:33just learn from books, but simply what
43:36we acquire through our experiences,
43:38through our experiences and so on. Yes,
43:43all of that sticks with us somehow, but
43:46that's not rational knowledge, yes,
43:48it's a bit of deep knowledge, also
43:50called test knowledge, that's the
43:52silent knowledge. Um, and there are
43:56also methods for accessing your
43:58experiential knowledge and your
43:59intuition and using them consciously.
44:03And you can train all ten of these
44:05skills. Um, we just don't learn that in
44:09school, nor in university or in
44:11training. Um, and so far these have
44:14always been soft skills, right? Um, and
44:17these soft skills are now becoming the
44:18new hard skills, because
44:21>> AI is replacing other things that used
44:23to be hard skills, like being an Excel
44:26professional or PowerPoint professional
44:29or someone who can memorize lots of
44:31paragraphs and rules and so on. Um,
44:35these hard skills are no longer in
44:37demand, these new skills, which are not
44:40new. Um, but so to speak, these are
44:44becoming the new valuable skills,
44:46because they can only be performed by
44:49humans. AI can only emit a little bit
44:52of this, but it can't come close. So,
44:56these are our islands, so to speak, yes
44:58, that we perhaps don't need to save
45:00ourselves from, but that we should
45:02perhaps focus on.
45:04>> Mhm. And Professor Moing, in 2035, 2040
45:07, where would you rather go? In a
45:11courtroom where an AI makes a ruling,
45:14or the judge, or in a hospital, would
45:17you rather ask the AI or the chief
45:20medical officer?
45:26>> Who would you trust more? If it's the
45:29court, then the judge as a human being,
45:34everything that comes before that, all
45:36the correspondence and so on and so
45:38forth: AI can do that, it can do that
45:40almost just as well. In medicine, the
45:45doctor who is supported by AI. That's
45:53also an interesting area. Sure, AI can
45:56better evaluate data and also do MAT or
46:00whatever. But we know, a diagnosis or
46:04an illness, whatever, right? It's not
46:08always one-dimensional, it also depends
46:10on what kind of life circumstances the
46:12person has. Are there any other things?
46:15Does he or she do sports? Does he or
46:16she not do sports? Has he ever had a
46:18little number here or there? What is
46:20the mental state of this person and so
46:23on? Of course, an AI can get to the
46:25point and make a diagnosis much faster,
46:28but there is much more involved and
46:30humans can do that better.
46:32>> Yes, that was really exciting. I am
46:35also wondering, when we meet again in
46:3720 or 35, Professor Moring, whether we
46:39will then let our AI agents communicate
46:42with each other and whether the visual
46:44aspect will be just as good or perhaps
46:46even better. Yes. Yes. Or that we will
46:50somehow be connected to our brain in
46:52advance, so that we can control it in
46:55such a way that in the end it will be
46:57us. Nobody knows.
47:00>> That was really exciting. Also, dear
47:02viewers and Z Z Z Z viewers, thank you
47:04very much for joining us again for your
47:06time and trust and please don't forget
47:09to subscribe to the YouTube channel
47:11when you watch us. That way you won't
47:13miss any exciting talks in the future
47:16and that will help us immensely with
47:18the algorithm. And don't forget, it's
47:20free of charge, of course. It always
47:23sounds like you have to transfer € 10
47:25to me every month. Unfortunately,
47:27that's not the case. Of course, I could
47:29think about it. I would have certainly
47:33earned it, but let's leave it at that.
47:36In that sense, you're doing well. Many,
47:38many thanks again, Professor Moing.
47:40You're very
47:41>> welcome. Thank you very much.

About this transcript

This page contains the full transcript of KI-Experte warnt: „Was 2027 auf uns zukommt, unterschätzen alle!“ | Prof. Dr. Moring by Clasen Talk, generated from the public captions YouTube serves with the video. The transcript has 7,431 words across 1,124 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.