Latest / Rewire To Rise– Beyond The Mindset / AI & The Human Future | Episode 2: AI & Education (Live Panel)
Transcript
- 0:0099% of people today use AI like a toy.
- 0:04Write me an e-mail, generate a funny picture.
- 0:06They think they're touching the future but it's an illusion.
- 0:10The real game begins when you stop asking questions and start
- 0:14building systems. AI isn't a search engine
- 0:17anymore, it's your high speed copilot for total automation.
- 0:21The market is rewriting the rules right now.
- 0:24Remember this, AI won't replace humans.
- 0:27Humans will be replaced by other humans using AI.
- 0:31When you delegate the chaos to algorithms, you free up time for
- 0:35strategy and that is what makes you irreplaceable.
- 0:38Stop being just a consumer of technology and become the
- 0:41architect of your own value. The world is evolving and the
- 0:45future belongs to those who adapt and leave.
- 0:48You have the tools, you have the vision.
- 0:50The only question is, Are you ready to?
- 0:51Level up. Here we are.
- 0:57Let's wait a couple of seconds to make sure that everyone is
- 1:02here. Hello everyone.
- 1:36I'm Mirella Bocanet. As a brand strategist and
- 1:40certified Cognitive Coach, my mission is helping leaders
- 1:43navigate the intersection of human intelligence and the
- 1:47future of technology. Today, we are moving beyond the
- 1:51surface of AI to discuss AI and the human future.
- 1:55Specifically focused on education, we will explore
- 1:59whether we are augmenting our minds or simply replacing
- 2:03thoughts. Before I introduce our first
- 2:05specialist to quick logistics, scan the QR code you see on the
- 2:10screen to register. This ensures you receive our
- 2:14goodie bag directly in your inbox.
- 2:17You can also find the registration link in the
- 2:20comments and the event description.
- 2:24Let's dive into our first pillar of future literacy.
- 2:28Our first guest guest is Hello Welcome.
- 2:38Hello, Mirela. Thank you.
- 2:41Well, the stage is yours. Great.
- 2:45Thanks. On.
- 2:47My screen and get my presentation.
- 3:01So is the presentation on it seems to be So thank you for the
- 3:14invitation and thank you everyone for, for your time.
- 3:18I've been looking so much forward to, to, to this event
- 3:21because there is so much to share, not only from me, but
- 3:25well, first start from me. And The thing is that AI effects
- 3:31every part of our society. Today we focus on the
- 3:35educational system and it indeed affects the the educational
- 3:40system. It affects for start the
- 3:43students and even starting at an early age, it affects the school
- 3:48system. And sometimes they actually
- 3:51might end up in situations that they cannot see through all the
- 3:56details and all the consequences.
- 3:58And that's the topic for my presentation.
- 4:01I've chosen to call it the digital landlord.
- 4:06The thing is that if you give or if you want to use a service,
- 4:14you can either pay with money or you can pay with data.
- 4:18Sometimes you'll pay with a combination and most systems
- 4:24used in schools, especially when talking about AI is really a
- 4:29combination because The thing is that a large pool of structured
- 4:35data is very valuable. It is valuable to the supplier
- 4:41that might differ from one country to another.
- 4:43And I'll get to more details in in, in that in the Danish
- 4:48market, that's where I'm located, it will often be
- 4:51Microsoft. In other markets, it might be
- 4:53other local suppliers and there will be different regulations.
- 4:56And I'll get towards all of that.
- 4:59In the case of the Danish system, in the case of using
- 5:03Microsoft, there is a certain payment in money, but there is
- 5:09also a payment in data. Because once you give away data,
- 5:14once that data is structured well, well, the system can serve
- 5:17you better, but it's not the entire thing.
- 5:21You basically pay a rent. You give them a way more and
- 5:25more data. The more data you get to give
- 5:27away, the more you are, the better you're served, but the
- 5:31more you also depend on that digital landlord.
- 5:37So we really don't own our data. We give our data, we hope that
- 5:42they treat it well. We hope that they serve us well.
- 5:45And there are a lot of things to consider in the in, in that
- 5:50situation. And it might not only be about
- 5:52Microsoft, it could just as well, it'd be about Google or
- 5:55Open AI and a lot of other players in this market.
- 6:01And of course, I'll get to more details in the Danish market.
- 6:08That's where I'm located. Much of it is really based on
- 6:13the public sector, and the public sector in this case pays
- 6:20for the suppliers, mainly Microsoft.
- 6:24Four and one quarter billion Danish corner per year.
- 6:30One euro would be 7 1/2 Danish corner, so it would be roughly
- 6:35half a billion euros. Whether that's a lot of money or
- 6:40whether that is not all that much money, I guess that depends
- 6:44on the perspective, but I guess it's quite an amount.
- 6:47So we do pay with money in order to protect data, but we also pay
- 6:51with data because the system should serve us well and we need
- 6:58not pay it to Microsoft in the Danish market.
- 7:01Most of it is paid to Microsoft. And of course they have expenses
- 7:06for building data centers and running services.
- 7:08So we're so I guess that as a starting point, well, it's a,
- 7:13it's a fair deal. We could in reality give it to
- 7:16other companies. I've chosen in this case that it
- 7:19could be Google, it could be open and eye, it could be a lot
- 7:24of different companies in the Danish market and it is
- 7:27different in other markets. I'll get back to that.
- 7:31But in the Danish one, it is often American suppliers because
- 7:35we don't have big and strong local Danish suppliers for the
- 7:42purpose that is different in other European countries.
- 7:45And, and I'll get back to that. This is not just my take on it.
- 7:50I have done my research very thoroughly.
- 7:55I have linked to sources. In this case I have linked to a
- 8:01Gartner report and I will link to more sources both in my
- 8:04presentation and in the material in the goody bag that you will
- 8:08all get access to after this. So yes, there is a payment, yes
- 8:13there is a data protection, and yes, there is more to the story
- 8:18because the story starts at a very early age.
- 8:23A student will often start in school at the age of 6 or 7 or
- 8:298. It depends a little on the
- 8:31country and it will continue for many years.
- 8:34And in the case that you could be served by Microsoft, which is
- 8:39the case in the Danish market, they are the biggest player in
- 8:43business systems. So it might actually be that you
- 8:46use the systems from Microsoft at a later age.
- 8:51They are the owner of LinkedIn, which is a big platform for job
- 8:55seekers and for hiring agents. And if they are tempted to reuse
- 9:01some of that data, and I guess the temptation is quite big, it
- 9:05could actually mean that what you do at the age of 6, what you
- 9:09do at the age of 12, what you do at the age of 18, and whatever
- 9:13is your path in the educational system will actually define how
- 9:18you will be seen as a job seeker late in life, how you will be
- 9:22seen as an employee late in life.
- 9:25There are possible regulations. And I hope that those
- 9:29regulations on place because yes, it is nice to be served
- 9:33well and these systems do serve you well.
- 9:36But well, how, if they serve a future employer will and that is
- 9:44not for the benefit of you. Once again, I have a source
- 9:48because I mean, if it was only my point of view, well, it would
- 9:53only be my point of view. And in many cases it is really
- 9:59so detailed. The systems work in a way that
- 10:04you can ask questions, you can get answers, but it will detect
- 10:08how you ask the question. It will detect if you do a
- 10:11follow up question, it will detect if you write perfectly
- 10:16well. If you don't write perfectly
- 10:18well, it will detect if you wait for a second or for 10 seconds
- 10:23before you are asked the follow up question, it might detect
- 10:27every handover that you do to the system.
- 10:30It might detect every action that a teacher takes.
- 10:37So it is really creating what I would call a digital twin.
- 10:43If they are in accordance with current regulations, and I hope
- 10:49they are, I really also believe that they are.
- 10:52They can keep that data for a certain time, not forever, but
- 10:56they can anonymize that and they can start building patterns and
- 11:01you will fit into that pattern. And a pattern that is not
- 11:04transferred directly into a very specific person, well, that can
- 11:09be kept for a long time. So it might be that at a later
- 11:13age when you search for your first job, you're not identified
- 11:17directly as a person because that would be illegal, but you
- 11:21are identified as a pattern. And maybe that pattern has so
- 11:24many characteristics to it that, well, you are really so well
- 11:29identified. Many of us might know that this
- 11:32has for a long time been the case with the with Google that
- 11:35has been independently of the the educational system.
- 11:40But if you use Google, it knows the mails that you send.
- 11:43If you use Gmail, it knows where you goes.
- 11:45If you use Google Maps, it knows what you search for.
- 11:49If you use the search engine, it knows what you're interested in.
- 11:52If you use YouTube. It is such a big ecosystem.
- 11:56And of course, Microsoft would so badly like to have an equally
- 12:02big ecosystem as Google and an equally detailed and valuable
- 12:07data source, because that's part of their business.
- 12:10It might be for the good of you, it probably is in some
- 12:13situations, but it might not only be for the good of you.
- 12:16So that's really a lot to consider and the people in
- 12:20charge might fully understand this.
- 12:23Sometimes I am in doubt. I mean, for me, I've done a lot
- 12:27of research. I know a lot about AI, but it
- 12:31it, it actually turned out to be.
- 12:35Well, let's start by calling it interesting, but sometimes also
- 12:40scary and sometimes overwhelming.
- 12:42There's a lot to the story, and because there is a lot to the
- 12:49story, and because there are many countries in the big world,
- 12:53because each country works in its own way, I have, for start,
- 12:58identified some big European countries that act differently
- 13:03than what we do in Denmark, where I'm located.
- 13:07Denmark isn't such a big country.
- 13:09It is a very, very digital country for the good and for the
- 13:13not all the way so good. In the case of Italy, there is
- 13:19actually a mandatory data sovereignty, so you're not
- 13:24allowed to use big American services like where Denmark is
- 13:30doing. It is called the Minerva
- 13:32project. It is backed by law and and
- 13:37there is a focus on making sure that you actually own your own
- 13:42data. And I think there is a lot of
- 13:44value to owning your own data. How that is handled, what
- 13:48suppliers, what will be the continuation of this thing,
- 13:52whether the big players might also get their hands on your
- 13:58data anyways? Well, that's the continuation of
- 14:01this. I know from organizations that I
- 14:04serve. I serve for instance, the Danish
- 14:06police, I shouldn't go into details.
- 14:09I serve the Danish military. I shouldn't go into details.
- 14:12I must serve a lot of organizations in Denmark,
- 14:15including the the authorities and the ministries and so on.
- 14:20And I know that they do have systems and a lot of good
- 14:24thoughts have been thought about those systems, but they're also
- 14:27private people doing private things and sometimes using
- 14:30services from Google and Microsoft and other players.
- 14:33And I'm quite sure it's the same thing in Italy.
- 14:35So of course, if you do a very good system for the educational
- 14:40purpose, but the students and the teachers and other players
- 14:44also use American service, well, they have legislation in place.
- 14:48And then there is more to the story.
- 14:52In the case of France, they have their local and very big and
- 14:56strong AI company, Mistral and and it is a keystone in the
- 15:03French approach. Mistral is a European country,
- 15:08It is or a European company, it is in a European country, it is
- 15:14under the EU legislation. But one of the biggest owners,
- 15:19one of the biggest investors is Peter Thiel, an American who
- 15:24owns the the surveillance company Palancia and who is one
- 15:30of the strong supporters of President Strong of our
- 15:33president there, Trump. So yes, it is a French company.
- 15:37Yes, it is a European country. Yes, it is EU legislation.
- 15:43And yes, there is at least some American influence to it.
- 15:48It might not be a problem, but it's worth a thought.
- 15:51In the case of Germany, SAP is a big local company and they have
- 15:56their own services. In the case of Finland, they
- 15:59have their ways. It is quite interesting, at
- 16:02least from my point of view, that these are EU countries.
- 16:07They are under the same legislation, but they make very
- 16:11different choices. And I cannot say that one is
- 16:14right and the other one is wrong.
- 16:16But I think it's worth a thought, especially as a Dane in
- 16:20a country where we use a lot of American services from
- 16:24Microsoft. There definitely advantages and
- 16:28disadvantages to every approach. And I guess it is at the end of
- 16:34the day a political decision. And I hope that the politicians
- 16:39have the qualifications to make relevant choice.
- 16:42Because in this case it might affect A6 year old, it might
- 16:46affect a 12 year old, it might affect the job search
- 16:51afterwards, it might affect the work life afterwards.
- 16:54And I hope that they serve as well.
- 16:57And sometimes I have my thoughts whether there is more to the
- 17:00story. But then again, it's not my
- 17:02decision. I try to present how it's all
- 17:06made-up, and I hope that the proper persons will make proper
- 17:12decisions. If you use free tools and some
- 17:20do well, you pay with data. And if you pay with personal
- 17:24data, and if you pay with a lot of personal data, including how
- 17:28you spell, including how you write, including how you talk,
- 17:32including how others see what you do, which is the case in the
- 17:36educational system, the consequences can be fairly big.
- 17:41From my part, I've been working with online marketing for many
- 17:44years, which means that I've been using the Google tools for
- 17:47many years. And my very basic understanding
- 17:51and the short story was really that I gave away data and it was
- 17:55used for tailor made advertising.
- 17:57I can live with that trade off, but this is much more.
- 18:02It is really about your life. It is at the end of the day
- 18:05about your thoughts, it is about your career, it is about your
- 18:09possibilities in life. And I hope that we can protect
- 18:13data. I hope that we don't give away
- 18:16too much data for for other players having a totally
- 18:21different approach to it. And there is legislation in
- 18:24place and I hope that we can make relevant legislation in
- 18:30order to, to support all of this properly.
- 18:36There are different approaches even in Denmark.
- 18:40Denmark is to a very large extent Microsoft country for the
- 18:46good and for the not always so good.
- 18:49I mean, Microsoft has an office in in Denmark.
- 18:52Google doesn't and all the big American tech companies don't.
- 18:56So we're so Microsoft is present in Denmark, but there are other
- 19:03approaches. In this case it is about a
- 19:06system called Leo, it is about a system called school GBT.
- 19:11And The thing is that those systems are made for education,
- 19:16which means that when you use the Microsoft service Copilot,
- 19:21it will deliver your give you the answer right away, but the
- 19:26school systems refuse to give the answer right away because
- 19:30you should be part of the process.
- 19:31You should do your own thinking. The school systems will ask
- 19:36guiding questions. They will preserve this, what I
- 19:40call the desirable difficulty. I know that just will talk much
- 19:44more more about it because he isn't the school system on a
- 19:48daily basis. And it means that there are
- 19:52tools made specifically for learning and there are tools
- 19:56made specifically basically for business efficiency.
- 20:00And I hope that we can combine these two approaches.
- 20:04It is relevant to have a knowledge about systems that are
- 20:07used in business life. Most will go there after the
- 20:11education, but I guess that education has its own dynamics
- 20:17and I'm happy that after. My presentation, other qualified
- 20:22and very skilled people will give their approach on it
- 20:25because it's really important to make relevant choices, to choose
- 20:29relevant systems, to make processes relevant for learning,
- 20:34which is not the same as business processes for
- 20:37efficiency. There are regulations in place.
- 20:45The short story is about the AI Act, which is about protecting
- 20:51data, it is about protecting people and it is about once you
- 20:56give away data, how long can they be stored, how much can
- 20:59they they be be tied to a specific person and so on.
- 21:03But then again, if these big companies may make the digital
- 21:07twin, which is not identifiable as a person, but is very, very
- 21:13close to an identifiable person, well, there is still something
- 21:17to talk, talk about. I will not go into details about
- 21:21the, the legislation also because of time constraints, but
- 21:24I'm happy that there is legislation in place.
- 21:27I hope that we manage and I hope that we don't leave too many
- 21:31loopholes for, for, for the, the, the IT players in this
- 21:36market. So from my point of view, you
- 21:42should ask these very essential questions.
- 21:46Who actually owns the data? Is there actually a human
- 21:50oversight? And is there an exit strategy?
- 21:54And the exit strategy might actually be challenging.
- 21:57If you give away data and then more data and then even more
- 22:00data and it serves you better and better and better.
- 22:03And the supplier is so interested in getting data
- 22:06because it also improves their services.
- 22:09Well, it might actually be challenging to to to exit from
- 22:13that. I can see from many Danish
- 22:15companies that they use Microsoft because it is so
- 22:18efficient and well connected and the price point seems to be OK
- 22:22and they basically put in in handcuffs.
- 22:26I cannot say that it's necessarily a bad thing or a
- 22:29good thing. I can just say that it's a topic
- 22:32and it can be very, very expensive and very, very
- 22:35challenging to go to a different supplier afterwards.
- 22:39If you use Google services or Apple Services as a private
- 22:42person, you probably know from that point of view.
- 22:46So I guess this is my part of, of this session, at least for
- 22:51start. There will be the questions and
- 22:53answers afterwards. There will be very skilled
- 22:56people presenting after me and I look forward to, to hearing
- 22:59those additional perspectives. There's also the goody bag
- 23:03afterwards. And in the goody bags, there are
- 23:05these links. There is also a document telling
- 23:07a little more about it. But then again, I guess that my
- 23:10time is up by now. So thank you for, for, for your
- 23:14time. I'll pass on to the next one,
- 23:17and I'll look forward to it to the dialogue afterwards.
- 23:20Thank you, Cara. Thank you so very much.
- 23:31So then our next speaker, it's chosen.
- 23:35Not anymore. Hello.
- 23:46Hello here we. Are welcome the wrong button.
- 23:51Thank you for being here again. Yes.
- 23:56Floor is yours. Thank you and I will take to my
- 24:02presentation right away. Yeah.
- 24:12And here we are. I think yes, yeah, I will speak
- 24:20about yeah, the nervous system and the right to trust what we
- 24:31learn because what we are seeing on the education, what we are
- 24:37learning, do we can we trust what the, what the AI tell us
- 24:44what we what we should, should learn.
- 24:49And we, we, we don't always be sure what, what, what the how,
- 24:59how, how it can, how, how it can be trust.
- 25:03So what, what, what changes when we when we learn, when we learn
- 25:08with the with the AI before the AI there was there was books,
- 25:20there was pencils, there was teeth.
- 25:27Most of the in schools they, they, they use AI.
- 25:35And when we when we use the AI in in school, we the, the AI can
- 25:45be very, very also common confidence to know they the AI
- 25:55is very authority. So, so the students can they can
- 26:04become more lower because what, what deliver they are not so, so
- 26:09perfect like the machine can can deliver it.
- 26:13So they can, they can feel lower because, because they, they feel
- 26:19the pressure to do, do the best the student can do.
- 26:28So you're here. I already talked about the
- 26:31perfect answer. It's already there in the
- 26:34machine. So you keep asking, asking the
- 26:42machine how, how I do this. And it just keeps answering how
- 26:46you do this. But you can't think right,
- 26:49because the friction isn't there.
- 26:53And it can be very, very difficult to, to learn something
- 27:00if you don't have it in your body after, afterwards in your,
- 27:03in your mind. You can have something right
- 27:06away in your, in your, in your, in your, in your, in your brain,
- 27:10but it's not certain. It keeps the the knowledge in
- 27:20you. So it's like fog in the wind.
- 27:24And so you go back to the machine and ask the same
- 27:26questions again. Yes.
- 27:33So when something can sound true without being true, it's a very,
- 27:39it's a, it's a very. How to say?
- 27:49You, you, you, you, you have it all in your in your mind and you
- 27:53tell like me what what you, what you know and and sometimes that
- 28:04what you know from which isn't right on.
- 28:14That's, that's can be a problem in, in, in some way.
- 28:22So what happens inside the student?
- 28:26We, we, we've got the dysregulated nervous system that
- 28:30learns less. Well, like I talked about, it's
- 28:35don't stay inside in inside the the mind.
- 28:41So good as, as you have to to learn it.
- 28:51So yeah, it's, it's very yeah, you get you get it all and you
- 29:04got it back afterwards if you don't know what I'm talking
- 29:07about. Yeah.
- 29:10So if you're sitting there and you you teach, you learn and
- 29:17you're and you work with the AI, sometimes you have to take a
- 29:25break and regulate before you continue.
- 29:30So you don't think right again. Here you'll see a free
- 29:42exercises. I would tell name it.
- 29:48What I do is always I, I keep my feet on the ground and breathe
- 29:54more slowly and look away. What I'm what I'm looking at
- 30:03does it where is it, where this come from?
- 30:08My what to what to the answer missing and what argue argue
- 30:16against it, you know, and. Or do I have to go away to to
- 30:26the reality again, away from the screen, talk with some some
- 30:31other persons with the teacher, take a book read or something
- 30:36like that? Yes.
- 30:40Oh, so yeah. What are we going to do with the
- 31:11the children who isn't developed enough to know what what they
- 31:19are seeing? That is my my kind of work, what
- 31:25I'm working with to to take care of this kind of.
- 31:38It's kind of education with the with the AI.
- 31:43So with my humble, humble presentations, I think I will
- 32:00give the word right to Mark again.
- 32:03To Mark. Well, I, I say thank you for
- 32:05this. I hope you can't understand what
- 32:08I'm saying, but there's something good in the goodie bag
- 32:12for you, yes? Good job, Tristan.
- 32:16Thank you. Thank.
- 32:17You. OK.
- 32:22So the next speaker will be Mark.
- 32:27Hello, Mark, welcome. Thank you.
- 32:29Thank you. So have you here?
- 32:31So happy to have you here. Great to be here again.
- 32:35Yeah, the floor is yours. Thank you, I'll just I can get
- 32:40the presentation working here. Yeah.
- 32:46I think that's it, yes. I'd like to take over a bit
- 33:00after where Torsten left off. Torsten talked a bit about what
- 33:05it does to the psyche to to deal with AI, to work with AI.
- 33:11And I'd like to take it a bit further and maybe look into a
- 33:15bit about what we can do about it a bit further than what
- 33:19Torsten was talking about. I'd like to start with with a
- 33:23question, and I'd like to sit and let you sit with it for a
- 33:27moment before I actually answer it.
- 33:30What does AI do to the thinking you never even know you're
- 33:35doing? Not the thinking you outsource
- 33:39deliberately, not the essay you ask AI to write for you, but the
- 33:44slower hearted thinking, the kind that happens when you don't
- 33:47know the answer yet, when you're wrestling with a problem you
- 33:51haven't fully formed. That's what I'd like to explore
- 33:55today. Not what AI can do for learners,
- 33:58but what it might be doing to learning itself.
- 34:11About me. My name is Mark Sinclair
- 34:13Fleeton, and I'm the Editor in Chief of AEP, and I've written a
- 34:18book about the choices we have to make when it comes to AI.
- 34:24And my work lives at the intersection of AI, cognition,
- 34:27and human judgement, which is exactly where today's
- 34:31conversation sets. And these are four things I'd
- 34:36like to cover with you in the next few minutes.
- 34:39What AI is actually doing to the thinking we never noticed we're
- 34:42doing. The cognitive science that
- 34:45explains why the judgement we were supposed to be building
- 34:49along the way. And I'd like to finish off with
- 34:53a more concrete method that you can use from now on.
- 35:03Now, Torsten has just taken you somewhere important.
- 35:07He's shown you what this technology does to the nervous
- 35:10system, to your relationship with uncertainty, with trust,
- 35:16with your own sense of reality. In a sense, those kind of
- 35:21thoughts can make you a bit uncomfortable, perhaps even a
- 35:25little unsettled. And I would say that is a good
- 35:31thing, because here's the first thing that I'd like to say that
- 35:36that feeling is not a problem. It's data, and data is something
- 35:41we can work with. My job in the next few minutes
- 35:44is to take you from feeling to thinking, from awareness to
- 35:50decision, from knowing the risk to having a method for staying
- 35:55in control of your own mind. And that's what I mean by mental
- 36:05sovereignty. And it's the only answer I know
- 36:09at the moment to to the challenges we are facing.
- 36:19I'd like to something start with something that sounds simple,
- 36:22but that has quite profound implications.
- 36:26The struggle is not a side effect of learning, it is the
- 36:29learning. And to illustrate this, I'd like
- 36:34to bring out three concepts from cognitive science that explain
- 36:38why AI is a particular kind of challenge in an in an
- 36:43educational setting and not just a general one.
- 36:48The first concept is what we call desirable difficult
- 36:53difficulty, I think. I think already Cora mentioned
- 36:57this in his presentation. It means that when learning
- 37:01feels harder, retention is actually deeper tasks that
- 37:05require you to retrieve, connect, and apply information
- 37:09rather than simply receive it. Build durable understanding.
- 37:13This is called desirable difficulty difficulty in the
- 37:17research literature. AI often removes desirable
- 37:21difficulty when it's used too early in the learning process.
- 37:27The next concept is the generation effect.
- 37:31Information you generate yourself is remembered far
- 37:35better than information you were given or that you generate.
- 37:39In this case, when a student writes their own argument,
- 37:44however imperfectly, however haltingly, they own it when AI
- 37:49writes most of it for them, the risk becoming readers of they
- 37:53risk becoming readers of someone else's thinking rather than
- 37:56producers of their own. And readers tend to forget far
- 38:00more than writers do. 3rd concept I'd like to bring out is
- 38:06metacognitive blindness, and this is perhaps the most
- 38:09dangerous one. Students may become less likely
- 38:13to notice when they don't fully understand AI, as Torsten says,
- 38:18provides fluent, confident answers, even the questions the
- 38:21student hasn't fully formed. And if you never have to notice
- 38:25that you're confused, you'll never develop the capacity to
- 38:28resolve confusion yourself. Let me put the distinction as
- 38:43sharply as I can. In business, which is what we
- 38:47covered in Episode 1, AI can replace a task.
- 38:51A leader who outsources judgement can, at least in
- 38:54principle, be held accountable. They made a choice.
- 38:57They signed off on it. The accountability remains
- 39:01human, even if the thinking was partially dedicated.
- 39:05But in education, something categorically different is at
- 39:09stake. In education, AI can replace the
- 39:13formation of a mind if AI is used as substitution rather than
- 39:18support. The risk in education is deeper
- 39:21than in most workplace settings. In education, the risk is not
- 39:26only task substitution, but the weakening of the judgement
- 39:30formation we were supposed to be building.
- 39:34Those are not the same problem. A leader who uses AI poorly can
- 39:38be corrected. A student will never develop the
- 39:41capacity to think independently, has nothing to fall back on when
- 39:45the tools aren't there, No accumulated judgement, no
- 39:49pattern recognition, no tolerance for ambiguity.
- 39:54And that is the risk I think we're not taking seriously
- 39:57enough. Let me try to make this a bit
- 40:06more concrete. Let's take two students.
- 40:09Given the same assignment, write a 500 word argument for or
- 40:14against social media H limits. Very very contemporary subject,
- 40:23so to speak. Suddenly student A opens AI
- 40:30immediately puts in a front. Write a balanced argument for
- 40:33social media. H limits 500 words, year ten
- 40:36level reads it, makes minor edits, submits the issue.
- 40:42The essay that comes out is well structured.
- 40:44The argument is coherent, the vocabulary is appropriate.
- 40:48But at no point, not once, that the student decide what they
- 40:53actually think. Now student B spends 10 minutes
- 40:58writing a rough first paragraph, takes a position.
- 41:01It's imperfect. Some of the arguments are weak,
- 41:05but they're hers. Then she opens AI.
- 41:08I'm arguing for age limits. What's the strongest counter
- 41:11argument I haven't addressed? She revises.
- 41:15She pushes back again. The essay that comes out is
- 41:18messier than student A's, but student B now has an opinion and
- 41:22the ability to defend it. Which of those students would
- 41:26you want making decisions in 10 years time?
- 41:39A grade can be used to measure what a student could do.
- 41:43Now we need to decide what should it actually measure.
- 41:48Can a student produce a well argued text?
- 41:51Well, AI can now do that convincingly and very fast.
- 41:56So if that's what we're assessing, we're assessing AI.
- 42:01Can a student identify the weak argument in someone else's work?
- 42:05That's harder to outsource. It requires actual judgement.
- 42:10Can a student explain, live and under pressure why they took the
- 42:14possession they did? This is much harder to delegate
- 42:18and much closer to the skill we actually care about.
- 42:23Can a student recognize when an AI generated answer is wrong and
- 42:27say why? That requires genuine mastery of
- 42:30the domain. You can't fake it.
- 42:34These aren't rhetorical questions, they're a design
- 42:38brief. The assessments that will
- 42:40survive the next decade are the ones built around capabilities
- 42:44AI cannot fake. Now, I've been talking a lot
- 42:52about judgement, and I think I should define what I mean when I
- 42:56use the word, because the word often gets used very loosely in
- 43:03this context. I use judgement as the capacity
- 43:06to evaluate, design and take a position under under
- 43:10uncertainty, under pressure. What I mean is, judgement is not
- 43:15a personality trait, it's not something you're born with.
- 43:19It's a skill, and like every skill, it's built through
- 43:23repeated exposure to difficult situations.
- 43:29What I'm showing here, what I would say we're like 3 building
- 43:33blocks of judgement, 3 parts of what judgement is, and these are
- 43:39three parts that AI tend to bypass.
- 43:43First, pattern recognition across many attempts.
- 43:48Judgement accumulates from experience of being wrong,
- 43:51revising and trying again. You cannot borrow this from AI.
- 43:56You cannot inherit it from a fluent answer.
- 43:59It has to be earned. Second, the ability to tolerate
- 44:04ambiguity. A student who always gets a
- 44:08clean AI answer never learns to sit with uncertainty.
- 44:11But most real decisions in work, in relationships, in life live
- 44:17in exactly that uncertain space. If you've never practiced
- 44:21tolerating it, you will be lost when it matters.
- 44:26And thirdly, knowing the limits of your own knowledge, the most
- 44:30important metacognitive skill? Recognizing what you don't know.
- 44:37AI models don't demonstrate this.
- 44:39They generate confident text whether they know the answer or
- 44:42not. Students who rely on them stop
- 44:45practicing epistemic humility, and that is a serious quiet
- 44:50loss. Now, what we're looking into is
- 45:03more precisely what I would set up as two different kinds of
- 45:07futures. They're not far apart in in
- 45:11time, they could be the same classroom, the same students,
- 45:15the same generation, where you can find these traits.
- 45:19From the outside, they may look identical.
- 45:24The vast future is where we see a generation that uses AI.
- 45:30They prompt well. They know when to trust the
- 45:32output and when to question it. They use AI as a tool and remain
- 45:37in control of their own thinking.
- 45:39They can explain every conclusion.
- 45:41They put their name to it. Now the second future is a
- 45:46generation that kind of think without AI.
- 45:50They also produce fluent work, perhaps more of it, perhaps even
- 45:54faster. But if you ask them to defend it
- 45:56without the tool, they can't. They never built the judgement
- 46:01that makes AI meaningful. And this is the part that
- 46:04concerns me the most. They don't know what they've
- 46:08lost. The question is not which future
- 46:12AI will bring us. The question is which future we
- 46:15choose to design for. So a lot of fancy words, some of
- 46:26it you will find in explain in the goodie bag, some of the
- 46:31terms I've used here and and even some of the sources.
- 46:35But I'd like to give you something specific you can take
- 46:38away from this. It's something that you can
- 46:41actually do something with. How can you counter these
- 46:45dangers? And what I've set up is, is a
- 46:48method, it's a specific force that exercise you can use
- 46:51yourself or you can teach it today.
- 46:55I call it challenge the bot. The principle.
- 46:59The principle here is that is use AI to argue against your own
- 47:04position, to make your thinking stronger.
- 47:07The first step, form your own position first.
- 47:12Write your own argument, however rough, before opening any AI
- 47:16tool. This is non negotiable.
- 47:19The thinking must begin in your head.
- 47:22If you skip this step, this step, everything that follows is
- 47:25just editing someone else's work.
- 47:28You haven't thought you've just reviewed Step 2 trumped AI to
- 47:35oppose you. Now you can use this prompt or
- 47:40some variant of it, but my suggestion is to write, here is
- 47:44my argument. Then you paste your argument
- 47:47into the text. Give me the three strongest
- 47:50objections to this position. Where is my reasoning weakest
- 47:55now here. You're not asking AI to agree
- 47:57with you, you're asking it to stress test you.
- 48:02Step three, decide. Don't just revise.
- 48:06For each objection AI raises, do update your argument, find
- 48:11stronger evidence, or hold your position.
- 48:14Every choice must be a conscious decision, not an automatic edit.
- 48:19You need to know why you're changing something and why
- 48:22you're not. Now Step 4I call document your
- 48:28reasoning. Add one paragraph to your
- 48:31submission, your report, even to your thinking, which is the most
- 48:36important really. AI raised the objection that XI
- 48:41responded by Y. This single paragraph is the
- 48:45evidence of critical thinking. It's harder to fake because it
- 48:48makes the reasoning process visible.
- 48:51It requires you to have actually engaged with the counter
- 48:54argument, not just received it. This method works for students
- 48:59writing essays. It works for professionals
- 49:01making decisions. It works for anyone who wants to
- 49:04use AI as a thinking partner rather than a thinking
- 49:09replacement. Now, should you use this last
- 49:12step in a professional report that you prepared in a work
- 49:16situation? Well, maybe not.
- 49:18But it's, it's very important for you in this process that you
- 49:22actually consider why you have chosen to answer the way you
- 49:27did. And writing a paragraph like
- 49:30that can put you in that frame of mind of, of actually
- 49:34remembering that. Now, I'll leave you with the
- 49:41distinction I think matters most.
- 49:45We're not at risk of producing a generation that cannot use AI.
- 49:49We are at risk of reducing a generation that cannot think
- 49:53without it. Those are very different
- 49:56futures. 1 is a generation that uses a powerful tool wisely.
- 50:02The other is generation that has outsourced the capacity that
- 50:05makes wisdom possible in the 1st place.
- 50:08The choice is still ours, but only if we make it deliberately.
- 50:16Thank you. That's presentation from me.
- 50:23Perfect. Thank you so very much, Mark.
- 50:25I really appreciate it. We are going to the last but not
- 50:33least for our last speaker. Hello, just welcome.
- 50:40Hello, Marella. Hello.
- 50:42Hello, long time no see. Indeed, pleasure to be here
- 50:46today. OK, voila, the floor is yours.
- 51:00All yours. Yours.
- 51:02Thank you so much Marella. See you.
- 51:05Yep. So I'd like to follow up on
- 51:12these three great presentations with a presentation of my own on
- 51:16on the topic of today, AI and and learning.
- 51:20Since it's the first time that I'm I'm here on on this fine
- 51:24live cast, I'd like to start by introducing myself.
- 51:30I when I'm not doing podcasts like this, I help coordinate a
- 51:34couple of e-commerce degrees at the Business Academy Copenhagen.
- 51:39And in teaching e-commerce, I've had a front row seat to this
- 51:42fantastic development that has been artificial intelligence
- 51:47over the last couple of years before I decided to become a
- 51:52professor. Well, like most of the
- 51:54professors at the business Academy, I used to work for some
- 51:57large companies like Andal, BP, Austed, KPMG, and of course my
- 52:04favorite consulting company implement.
- 52:06So what I'd like to talk to about today's is sort of, you
- 52:11know, pick up on some of the things that Torsten, Mark and
- 52:16Core have already raised, but namely dive into sort of this
- 52:21thing that AI has become a separator of, of learning and
- 52:26doing. And also at the same time, AI
- 52:30has the potential to be both an enhancer and an inhibitor of
- 52:35learning. Adding to the those two sort of
- 52:39main points, I'd also like to go into the different perceptions
- 52:42that seem to be out there of AI usage and also talk about the
- 52:47future skill sets. What is the current consensus on
- 52:52what skills the future will require?
- 52:56But going in sort of to the first point, and I think this is
- 53:00this is a fairly obvious 1, is that a lot of learning
- 53:04institutions, they learn based on the premise that in order to
- 53:12produce something, you need to have acquired knowledge, You
- 53:15need to have process that knowledge.
- 53:17And only by sort of by giving students a task of applying
- 53:22something, well, you, you kind of force them into a learning
- 53:25process. And The thing is that that AI
- 53:28shifts that that spectrum round because all of a sudden doing
- 53:32something, creating something no longer really requires you to
- 53:36acquire a process knowledge. You can you can prompt your way
- 53:40into to having a finished product.
- 53:43And that means that the process of creation and the process of
- 53:47acquiring knowledge becomes 2 distinct fields, perhaps not
- 53:52entirely distinct, but they become more separated than they
- 53:56were before. And this is the challenge that
- 54:01that AI is, is basically raising in the educational sector.
- 54:07I mean, Co did a fantastic presentation on the, the state
- 54:11of sort of investments in the IT systems.
- 54:14But we have to remember that beyond that, there's also the
- 54:17whole question of shadow IT. So basically professors,
- 54:21teachers, students, pupils, they can all choose to use AI tools
- 54:28without the schools, the universities or the academies
- 54:31ever knowing about it. We we're a couple of years now
- 54:34past where we could sufficiently and effectively detect that.
- 54:38So basically the separation is here, like it or not.
- 54:43And this also gets into some sort of interesting things.
- 54:49A fairly obvious point, which a lot of people pointed out early
- 54:53on is that if you use AI to create something, well, you
- 54:58haven't processed information, so you, you can't really, you
- 55:03know, recall that. And one study that I think sort
- 55:05of hammers home that point is that when a, a group of, of
- 55:11students in the US, they were asked to use ChatGPT to write an
- 55:15essay. They were fairly satisfied with
- 55:18the essay. This, this looks great.
- 55:20They said they, they weren't entirely sure they agreed with
- 55:24all of the points in it. But the interesting part came
- 55:27when people said that people asked sort of, you know, can you
- 55:30quote something for me? Can you give me sort of repeat
- 55:33one of the lines that you're arguing?
- 55:34Did you write this or did you write that?
- 55:37And the vast majority, 80% of the students who'd written their
- 55:42essays using ChatGPT, they could not quote a single sentence.
- 55:49And that sort of is, you know, Hammersome the point sort of,
- 55:52you know, that that if you use the AI to create something,
- 55:55yeah, well, you know, something's been created, but
- 55:57the whole process that's typically has been associated
- 56:01with that creation process. Well, that hasn't happened.
- 56:03And that's sort of, you know, the one point from the study.
- 56:08The other point, which is also sort of an interesting aspect to
- 56:12raise here is that when we use AI, and indeed also when we use
- 56:17the Internet search engines and such, the topics that we seem to
- 56:23come back to, the quotes that we make, the sources that we refer
- 56:28to, they seem to coalesce. So if we ask humans to write an
- 56:31essay about, let's say, you know, argue against social media
- 56:36age limits, just to pick up on that example again, then humans,
- 56:41they will have a more diverse background of examples of
- 56:46different sources that they will quote.
- 56:48Then AI will AI will, you know, be a winner takes all in sense
- 56:53of who is quotable. And that's sort of, you know,
- 56:55not by, by ill will or bad design.
- 56:58That's simply the structure of the of, of, of how AI works.
- 57:04So I mean, and if that's sort of, you know, some of the things
- 57:08we have to be mindful of that. If we just use AI to create
- 57:11something, well, you know, we're, we're not going to get
- 57:13any learning. Well, that's the way that AI
- 57:15will potentially inhibit learning.
- 57:19And I like to put it this way that, you know, if we ever could
- 57:22force something, somebody to learn something, well, that's
- 57:24not true anymore. Now we can't, we can't force
- 57:28anybody to do anything in terms of learning.
- 57:30But there's a flip side to the story.
- 57:33And the flip side is that AI also has a great potential, a
- 57:38huge potential to enhance our learning if we practice with it,
- 57:42if we use it to resync, we use it to qualify.
- 57:44And I think Mark gave a great example here of, of challenging
- 57:49our view how that works. And that's sort of the way that
- 57:54AI can enhance all those things if we choose to let it do so.
- 57:59And I think sort of the future that we're looking into is all
- 58:02of a sudden students, they will be faced with the choice of what
- 58:09topics would they like to prioritise?
- 58:12Where would they just like to have a summarised level of
- 58:15knowledge and what topics that the school system requires of
- 58:19you do they essentially consider to be irrelevant to them.
- 58:26And again, sort of, you know, this is one of those things
- 58:28where I think we also have to consider a little bit when we
- 58:32say the student, what do we mean?
- 58:35Because a student can be anything from a six year old boy
- 58:40beginning in grade 0 in the Danish public school sector or
- 58:44it can be a 25 year old PhD student, right?
- 58:48And I think sort of there's a lot of evidence emerging that
- 58:52this choice is ill made by younger students.
- 58:57So I don't know if the cut off point is 18 or if it's if it's
- 59:01higher or lower, but it's probably around there somehow.
- 59:05There's been a number of surveys that show that if you, you, you
- 59:08deploy large unstructured general AIS and say the Danish
- 59:14middle school or something like that, well, you know, the
- 59:17inhibitor effects are going to dominate.
- 59:21Whereas there are other studies suggesting that deployment of,
- 59:25of AI among university level students.
- 59:29Well, you know, the enhancer effects are definitely there and
- 59:33can be captured. But then again, sort of, you
- 59:36know, we're pushing back sort of, you know, what do you really
- 59:38want to learn? And a little side point on that,
- 59:43you know, what is what is, what is the learning?
- 59:46You know, if we become the masters of our own learning with
- 59:51all the stress of having that choice that that entails, then
- 59:56we also have to decide before we Start learning something.
- 1:00:00Well, what level of knowledge am I am I going for here?
- 1:00:05And I see a lot of AI tools offering, you know, to
- 1:00:08summarise. Offering to create quizzes based
- 1:00:12on say, the professor's PowerPoint slides.
- 1:00:15But that approach specifically if you, if you take and you know
- 1:00:18you use an AI and engage with it based on the professor's
- 1:00:22PowerPoint slides. But that's sort of assuming that
- 1:00:24all the relevant information you need to know are on the slides.
- 1:00:28What if all the relevant information was in the
- 1:00:31underlying book and the slides were simply a simplification of
- 1:00:35that? Then you're sort of getting a
- 1:00:38summary of a simplification. All of a sudden, you know, we're
- 1:00:41ending up with people saying, well, Mona Lisa is just a
- 1:00:45smiley. And I, I hope we can on this
- 1:00:47call all agree that, you know, it's, it's so much more really.
- 1:00:52So there's another thing here. One thing that I'd, I, I'd like
- 1:00:56to notice that there seems to be a new meaning of the word
- 1:01:01hallucination. If you go back to sort of pre AI
- 1:01:07years, you know, so 2020 then hallucination was a
- 1:01:13psychological term. It meant when sort of humans,
- 1:01:16they imagine something that wasn't there when we heard
- 1:01:19voices, you know, felt, you know, touches that, you know,
- 1:01:24didn't happen. And now sort of hallucination
- 1:01:28has sort of broadened up a little bit, at least it's used a
- 1:01:30lot in these years in association with General II,
- 1:01:34meaning when a General II comes out with an answer that's
- 1:01:38factually incorrect. And I just have to say sort of
- 1:01:40if this is the new meaning of hallucinations, well, I've seen
- 1:01:42humans hallucinate a lot during exams in my years as a
- 1:01:46professor. So we're sort of getting into
- 1:01:51the point where we're moving in and saying hallucination might
- 1:01:56be you're changing the meaning and sort of the way we're, we're
- 1:01:59using it. And I would, I would argue a
- 1:02:03little bit against sort of using that, that for humans, even
- 1:02:06though I, I just did that. There's another point of, of
- 1:02:09view to this. And that is sort of we are
- 1:02:12currently in some years where you can easily find negative
- 1:02:18reactions to AI usage. So I did a number of interviews
- 1:02:22with other professors at the place where I work.
- 1:02:26And one of them who's been anonymized here, told an
- 1:02:29interesting little story that, you know, when he conduct exams,
- 1:02:32he'd like to ask students, you know, how they used AI, you
- 1:02:36know, and if they used it sort of like Mark gave an example of,
- 1:02:40you know, where they are, you know, used it to to argue
- 1:02:43against their point. Well, you know, he would
- 1:02:45consider that to be a good way of using it.
- 1:02:48But then this guy also said, I have colleagues who view the
- 1:02:51usage of AI as a flaw and they will simply give a lower grade
- 1:02:56if AI has been used no matter how it's been used.
- 1:02:59And the guy I was interviewing said, so, you know, if I'm
- 1:03:02sitting, you know, and has one of these colleagues as my Co
- 1:03:04examiner, I'll simply refrain from asking any questions about
- 1:03:08AI usage because better not to stir the pot, right?
- 1:03:13So there's a study here from the UK which basically says that
- 1:03:19apparently, you know, the AI negativity seems to be quite
- 1:03:25commonplace in general, when people they are told that
- 1:03:30something has been created by AI, well, they seem to view the
- 1:03:37person doing that as less competent, less warm, less
- 1:03:41morally good, more lazy and less trustworthy.
- 1:03:46So all of those bad things. I mean it, it varies a little
- 1:03:49bit, you know, by what you're sort of, you know, you know
- 1:03:53what, what example you're giving, you know, if it's, if
- 1:03:56it's a love letter or a job application, it hits through
- 1:04:00very hard. If it's generating a grocery
- 1:04:02list, there still is a negative effect, but it's a lot lower and
- 1:04:06seems only to, you know, be be well, it's actually not even
- 1:04:11that negative. But The thing is sort of, you
- 1:04:15know, then and just to explain this slide here, because there
- 1:04:18was a little bit more to this study, because there was also
- 1:04:20the comparison, like if you'd had a buddy do it for you.
- 1:04:23So if I showed somebody a love letter and I say, well, you
- 1:04:27know, I had Mark write this love letter for me because he's great
- 1:04:30at writing love letters. Well, you know, and this thing
- 1:04:33is sort of that the negative views, they were the same as
- 1:04:38with if I said, you know, I used AI to write this love letter to
- 1:04:43my to my wife. And that, you know, it's perhaps
- 1:04:48where AI is today. If, if we used AI for a task,
- 1:04:54there's a good chance that a lot of other people will view it as
- 1:04:58if we had asked a different person to do it for us.
- 1:05:04And again, sort of, I'm not saying, you know, this is the
- 1:05:07right way to view. Yeah, I'm just saying that is
- 1:05:10the current view. So we have to be mindful about
- 1:05:13how we disclose our AI usage. And when I guess there's also
- 1:05:20sort of a thing that that that I observed.
- 1:05:22And this is sort of more or less just by my own little
- 1:05:24observation, but I've seen sort of, you know, this in a number
- 1:05:29of surveys though, that you can find sort of roughly 1/5 of of,
- 1:05:33of the student body describing themselves as having high AI
- 1:05:37literacy, high AI you choose. And you can find another 20%
- 1:05:39right now, you know, say, you know, I'm not using AI at all
- 1:05:43for my studies yet. When I talk to professors about
- 1:05:49sort of the quality of written assignments, well, they've all
- 1:05:52gotten better. All of the written reports that
- 1:05:55are being handed in and sort of so what's what's happening
- 1:05:59there? You know, they've also, it's
- 1:06:04also reported that there are some reports that now seem high
- 1:06:06quality, but you know, there are AI tells.
- 1:06:10So there's a specific way in which ChatGPT likes to write,
- 1:06:15you know, X is not this, it is that.
- 1:06:21And if you see a sentence structure like that, well then
- 1:06:24you know, that repeats in the, in, in the, in the essay or the
- 1:06:27assignment. Well, you know, that's, that's a
- 1:06:29likely AI tell different, different AI systems seem to
- 1:06:34have different tells. And you know, how accurate is
- 1:06:36the detection of this? Don't know.
- 1:06:38But the professors that I work with, they report that there are
- 1:06:42AI tells in some of the reports in a few of the reports, but all
- 1:06:46of them are high quality. So what's happening here and
- 1:06:50possible sort of, you know, explanation for this is that the
- 1:06:55two futures that Mark described are happening at the same time
- 1:07:01because everybody's making their own choice.
- 1:07:04Some students, they who are high in literacy, they know what AI
- 1:07:07is, they know what to represent. They use it for learning, they
- 1:07:09use it to come in and enhance their learning.
- 1:07:13The AI avoiding students, well, they're probably also fine if
- 1:07:16they have a different good learning technique.
- 1:07:18And then you have all the ones in the middle, the ones who
- 1:07:20don't really understand AI but still use it and in some cases
- 1:07:25even forget to use it in a way where you hide the AI tells.
- 1:07:31Now, I'd like to point out also that this debate about sort of
- 1:07:35how technology will impact learning, that is not a old
- 1:07:40debate. Plato recaps an argument between
- 1:07:45himself and Socrates in which Socrates went on a rant against
- 1:07:51books and Socrates argued, well, you know, books, you know, if
- 1:07:54you read them, they dull true understanding, true
- 1:07:57understanding. We can only get that from
- 1:07:58dialogue, from debate. And I guess Plato was a little
- 1:08:03bit like, well, yeah, but you know, I mean, if I write a book,
- 1:08:06you know, a lot more people can read it.
- 1:08:09Though he doesn't. He's not that specific on his
- 1:08:10pro arguments. He lets Socrates just have that
- 1:08:13point. And again, sort of, you know,
- 1:08:15this comes back to sort of what, what is our, our view of, of
- 1:08:21good learning. If we say that good learning is
- 1:08:24the ability to argue something on the spot, there might be a
- 1:08:28chance, you know, that, that the, the more you read, the less
- 1:08:33proficient you would become on that.
- 1:08:35You know, we'd all know the term street smart.
- 1:08:39And maybe that's a different kind than Professor Smart,
- 1:08:42right? So there's that as well, you
- 1:08:46know, but just to, you know, include this fantastic picture
- 1:08:50here by the Italian Renaissance painter Raphael.
- 1:08:55And also to point out sort of that that when we're sceptical
- 1:08:58of AI, well, you know, there's really nothing new under the sun
- 1:09:02here. There's also when we talk about
- 1:09:05AI, we run into the sort of the three big difficulties within
- 1:09:13AI. 1st, when we talk about sort of artificial intelligence, you
- 1:09:17know, well, we don't really have that much of a strong
- 1:09:22definition, common definition of what human intelligence is.
- 1:09:26And because we don't have a strong definition of that, we're
- 1:09:28struggling to define on sort of when we reach something that can
- 1:09:33match a human intelligence in any meaningful way.
- 1:09:37There's also sort of a diverse application of, of AIAI shows up
- 1:09:42in, in many different areas and it develops differently in
- 1:09:46different areas. And finally, we have a lack of,
- 1:09:50of, of common understanding of what AI is.
- 1:09:53So I've, I've given you some, some examples here because the
- 1:09:57point basically is that what we today would refer to as AI,
- 1:10:02maybe in a, in a few years, it will be so common people will
- 1:10:07not even think of it as being AI.
- 1:10:11I, I read a report recently which said that the, the most
- 1:10:16used AI in Denmark, that was facial recognition on phone, you
- 1:10:21know, to unlock your phone. And I was thinking, well, wait a
- 1:10:24second, there was facial recognition AI.
- 1:10:26And I guess, you know, by a lot of definitions it, it would be,
- 1:10:29but by some other definitions it wouldn't.
- 1:10:32And so I think, you know, there's there's that as well
- 1:10:35that we're talking here when we're talking about AI and
- 1:10:37learning, we're also talking about matching and moving target
- 1:10:42to something that has well, of which the definition is an
- 1:10:47organized chaos. And that makes the debate
- 1:10:50difficult. But you know, we should try
- 1:10:52nevertheless. And there I get to my final
- 1:10:55point, which is sort of, you know, what are the skill sets of
- 1:10:59the future? And I like to come back here to
- 1:11:04a couple of decades old book by some American professors,
- 1:11:11Trilling and Fidel, where they basically map that when people
- 1:11:15talk about sort of what skills we need to have for the future.
- 1:11:20Well, you know, what are those skills?
- 1:11:21And they, they identified some 12 skills and said, well, you
- 1:11:24know, these are the 21st century skills.
- 1:11:27And they, they put them into three main buckets.
- 1:11:32The first would be the ability to learn.
- 1:11:34The second would be the digital literacy.
- 1:11:38So basically knowing how technology works.
- 1:11:41And the final step would be what you could call life skills.
- 1:11:44And if, if, if we look at them, the point here is that these,
- 1:11:48these three buckets are equally important.
- 1:11:53So I think when we talk about the future of learning, and I
- 1:11:55still think this framework, 'cause we have to think about
- 1:11:59sort of learning needs to happen in all three areas.
- 1:12:02So we need to teach young people how to learn.
- 1:12:07We need to teach them how to understand the IT systems that
- 1:12:11are surrounding our everyday life.
- 1:12:13And we also need to teach them life skills.
- 1:12:15And there's a big debate, you know, about what should happen
- 1:12:19when and who should do the teaching, especially of life
- 1:12:22skills. You know, is that the job of the
- 1:12:23schools or the job of the of the family?
- 1:12:25That's a point of contention there.
- 1:12:28But I think I'll leave with that point that in the future, we
- 1:12:33need to have people who can both think critically and understand
- 1:12:38how AI works. And that would be my hand off.
- 1:12:44Back to you, Marella. Thank you.
- 1:12:47I really appreciate it very much.
- 1:12:50Nice job. Well, we kind of achieve arrive
- 1:12:54to the point of Q&A sessions, so I'm going to Add all our guest
- 1:12:59speakers. Hello again, Ebola, are we all
- 1:13:07here? Hello guys.
- 1:13:17We are going to start our Q&A session.
- 1:13:20Well, I'm kind of going to start with you.
- 1:13:24Since you have talked on how educational institutions are
- 1:13:28physically and structurally changing, please explore a tiny
- 1:13:33bit the training aspect. Thank you.
- 1:13:37Thank you for for for this follow up question because it's,
- 1:13:40it's so important. I mean, in the educational
- 1:13:43system, there are buildings and they're owned by whoever they're
- 1:13:46owned by and there are of our employees.
- 1:13:49And there are a lot of things that are not digital.
- 1:13:53But it seems that the digital heart is just growing and
- 1:13:57growing and growing, and it seems that it always depends on
- 1:14:03or usually depends on everyone except for the actual employees,
- 1:14:08which need not be a problem, but which might be a challenge.
- 1:14:14I guess that used already put good words to it and really
- 1:14:18everyone did. And as long as we are in charge,
- 1:14:21as long as it serves as well, as long as it's probably protected
- 1:14:25well, it need not be a problem. I mean, that can be reason for
- 1:14:29renting a car and that can be reasons for renting so many
- 1:14:32things. And when you have employees, I
- 1:14:34mean they are also private people with their private
- 1:14:37perspectives. The thing that I'm a bit worried
- 1:14:42about is that if we go so far down the path of the digital
- 1:14:48twin, which will not be deleted and which will affect students
- 1:14:53for basically many years ahead in their life.
- 1:14:57Well, we might have crossed the line.
- 1:14:59And I hope that we are served by politicians and other decision
- 1:15:06makers who are very aware of this and who succeed in
- 1:15:10navigating because it it need not be a problem, but it is so
- 1:15:15important that it's handled properly by people who actually
- 1:15:18understand the total of it all. Someone else wants to add
- 1:15:24something else. I guess that actually it is
- 1:15:28connected to to what you were presenting used so so you have
- 1:15:32probably some additional thoughts on this.
- 1:15:35Guys, if you don't speak, can you please mute yourself because
- 1:15:37otherwise we have the info and it's not quite very efficient.
- 1:15:41Thank you. Well, definitely, definitely so.
- 1:15:49So, yeah, I mean, I think that the education sector in Denmark,
- 1:15:55as I remarked to remember, Ella, well as in as in most other
- 1:16:00countries, I mean, the educational sector is benefits
- 1:16:05right now from being insulated to shocks in the sense that it
- 1:16:09is state owned, it's publicly financed, it's it's regulated.
- 1:16:14And so, you know, even if we have problems with the learning
- 1:16:19process, the machine continues to work.
- 1:16:23Yeah, I think there's a great point here also about, you know,
- 1:16:29these AI investments, if you were to be critical of them, you
- 1:16:33can ask Bill, you know, what's the business case?
- 1:16:35So, you know, we're spending roughly half a billion EUR, but
- 1:16:42what are we, what are we getting for it?
- 1:16:46And I guess, you know, if, if, if the majority of those monies
- 1:16:50is going, you know, to software like Microsoft Word, Outlook and
- 1:16:56so on, well, you know, it almost becomes like a license to
- 1:16:59operate, you know, but, but, but if, if you need to make
- 1:17:06additional investments in AI, you know, and you would start
- 1:17:10cutting down human stuff, that's what a private enterprise would,
- 1:17:14would do, right then, you know, how would you ever go about
- 1:17:18that? There's there's some structures
- 1:17:21in place that would make that move very difficult in most
- 1:17:25European countries. That's a strong point and I'm
- 1:17:29quite sure that Mark has a quite a good thought on on this.
- 1:17:34I know that in your book you have really been down this this
- 1:17:38path. Yeah.
- 1:17:44I guess what I'm most worried about is like we've talked about
- 1:17:48if we choose systems like you were explaining, we're using
- 1:17:54primarily you say Denmark as a Microsoft country.
- 1:17:58We're also in some sense an open AI country because we use chip,
- 1:18:01GPT and so on. And what we just talked about is
- 1:18:06how using AI influences the way we argue, the way we learn
- 1:18:14things. And if the, if the basis of what
- 1:18:19we're learning is actually systems which are not based in
- 1:18:23Denmark or in the Nordic countries, that is a huge
- 1:18:27problem. Because it, I mean, we've, we're
- 1:18:30already influenced of course in Denmark very much by the, by the
- 1:18:34US and, and used to be a lot more influenced by the UK.
- 1:18:39But this will certainly emphasize this development and
- 1:18:45that that could be a problem. I actually have a follow up
- 1:18:49question for for you, Mark. And I also have one for you,
- 1:18:52Torsten, but we'll take that next for, for, for you, Mark.
- 1:18:57I presented that other European countries have different
- 1:18:59approaches. I mean, France is basically
- 1:19:02using a French system. Germany is basically using a
- 1:19:05German system just to make the hours short.
- 1:19:08And I don't know whether that's a better approach, but it is a
- 1:19:11different approach from from the Danish one.
- 1:19:13And, and for you to prepare afterwards, Torsten, this is
- 1:19:16also affected actual humans. And I know you have a strong
- 1:19:20opinion on that. I sometimes hear you say that
- 1:19:23it's really a battle that we're about to lose.
- 1:19:27But but you can, you can do your thoughts and then Mark will
- 1:19:30probably have his say his say about whether a different
- 1:19:34European approach would be better than the Danish one.
- 1:19:39You remind guys that I also have some questions that I've
- 1:19:42gathered from social media for you for each one of you, a
- 1:19:46couple of questions just so you know.
- 1:19:47OK, perfect. Yeah, I think the, I think the
- 1:19:54Danish approach is lacking in a lot of ways.
- 1:19:58We are just building on systems that are not really like I said.
- 1:20:07Based on our way of thinking, our way of doing things.
- 1:20:11And I think we need to be much more aware of this for one
- 1:20:17thing. But the second, we need to
- 1:20:18develop for this. We don't necessarily need to
- 1:20:21develop Danish systems, although I'm in in favour of Danish
- 1:20:25systems, but I think we have to think more in Nordic terms
- 1:20:30because we have a lot of commonalities and the way we
- 1:20:33think and so on. So we could, we could, we could
- 1:20:36think in in Nordic systems. And I know that in both Sweden
- 1:20:39and Norway there are being developed bigger systems in
- 1:20:44this, but we should get on the way in Denmark as well.
- 1:20:49And I hope that that we can also have your perspective, Torsten.
- 1:20:52And then of course, we should also have the the the other
- 1:20:55questions from social media. How do you think about it,
- 1:20:58Torsten? Yeah, well, I have looked into
- 1:21:02it somewhere with the what it's called PBT and I have and I've
- 1:21:10looked everything true. Well, what it what it presents
- 1:21:16and I and I just. I think he has a connection
- 1:21:29problem, I don't. Need the answer and I need to
- 1:21:33how to, how to, how to, how to make the, the, the, the, the no,
- 1:21:44the up girl and to, to figure it out.
- 1:21:49And I asked, what are you doing? You, you, you, you just tell me
- 1:21:54the answer. You, you don't learn me
- 1:21:56anything. You don't teach me.
- 1:21:58So, so then it, it tried to tell me how to multiply, but the
- 1:22:04answer is still there. It's 56.
- 1:22:07I didn't learn anything. So, so, so I did something else
- 1:22:13and gave him some prompts in and it was different.
- 1:22:19So I'll ask you again. What is 8 multiplied 7 or 7
- 1:22:23multiplied 8? And then the AI asked me, Yeah.
- 1:22:30What? What do you, what do you know
- 1:22:32about it? Well, what's what result do you
- 1:22:35have? OK, then I have to think about
- 1:22:42it. What have I done here so so I
- 1:22:46can I can see it there. It's a it's a difference in in
- 1:22:49how how the the system teach the children and So what Mark always
- 1:22:55tell about the the the Nordic thing.
- 1:22:57It's it's about the culture. It's about how we how we
- 1:23:01perception of the things here in in our Nordic culture.
- 1:23:05It's nothing. It's not the same like the
- 1:23:07German or the English or something.
- 1:23:10It's also the a different. Yeah, there's surely cultural
- 1:23:21aspects to it. Knowledge thing about it, yes.
- 1:23:28If I could interject a small side comment to what you're
- 1:23:31saying to Austin, I think you know that within learning,
- 1:23:36there's, there's there's a theory of learning called the
- 1:23:38growth mindset versus a fixed mindset.
- 1:23:40So basically sort of, you know, are you learning in order to
- 1:23:43find out, do I have a natural innate ability for this?
- 1:23:47Why am I learning? Because I think I can better
- 1:23:49myself. And of course, sort of it's not
- 1:23:52like this, you know, people have one mindset.
- 1:23:54They have different mindsets across different areas.
- 1:23:56But I think there's also an impact here of AI in that it
- 1:23:59might, you know, discourage learning if people go about it
- 1:24:03with a fixed mindset. Because what will happen is that
- 1:24:06they will run up into the AI, you know, the perfect answer
- 1:24:09that you talked about, and they'll run into it and they'll
- 1:24:13say, oh, this machine is so good.
- 1:24:15I can never be that good, so why bother?
- 1:24:18I'm not good at it. And I think that's a risk as
- 1:24:21well. And the implication for this in
- 1:24:23the education system is that education moves even more in the
- 1:24:26direction of the role of the professor, being the
- 1:24:29cheerleader, the motivator, the guy that says, yes, you can do
- 1:24:32it. You know, is a trend that's
- 1:24:34always been there. But but, you know, I think AI
- 1:24:36also pushes it a little bit further.
- 1:24:37But Mirella, we had some questions from social media,
- 1:24:40didn't we? Yes, thank you.
- 1:24:44OK. So I have a couple of questions
- 1:24:46for it says it sounds like that how do we prevent a student's
- 1:24:53digital twin coating from becoming a permanent commercial
- 1:24:58profile for future recruitment? I hope that regulation will take
- 1:25:05care of it. But I know from experience from,
- 1:25:08from well many years of, of tech that it seems that we do the
- 1:25:13regulations, then the tech companies try to find the
- 1:25:16loopholes, we ask them to protect the data and sometimes
- 1:25:20they are unsuccessful in in doing it.
- 1:25:23And it is, it is really, it is really complicated.
- 1:25:26And I mean, it is not a black and white discussion because the
- 1:25:32digital twin can also serve as well.
- 1:25:34I mean, it is a possibility of, of a more tailor made
- 1:25:38experience. I, I see it from the, the, let's
- 1:25:41call it the traditional Google experience.
- 1:25:44I mean, I have chosen in my private life.
- 1:25:46I'm, I'm not a shooting any longer, but I've chosen in my
- 1:25:49private life to give a lot of data to, to Google because it
- 1:25:52serves me well when giving recommendations on YouTube, on
- 1:25:58other social media platforms. But the thing The thing is that
- 1:26:03if we don't know the rules of the game, we are not that well
- 1:26:06off. And I mean that the rules that I
- 1:26:09exchange data for tailor made advertising, well, I'm fine.
- 1:26:13But if I exchange data for basically my future career
- 1:26:18possibilities, I'm not into to that.
- 1:26:21And I really hope that politicians and other decisions
- 1:26:25maker will put the foot down in regards to, to, to that.
- 1:26:30Because in the case of Microsoft, they are also the
- 1:26:33owners of LinkedIn, which might affect your career.
- 1:26:36They are also the owner of business systems for sales and
- 1:26:40production and everything. And they have the possibility to
- 1:26:44connect it all together. But we really shouldn't allow
- 1:26:47them to do so. Yeah, there is another
- 1:26:50interesting question which you were actually talking about.
- 1:26:55Why are countries like Italy and France being more cautious than
- 1:26:59Denmark, and what are we sacrificing for convenience?
- 1:27:05My best guess, and I might not understand it all, but my best
- 1:27:09guess is that we really have different histories in, in, in
- 1:27:13these countries. I, I know that part of the, the,
- 1:27:17the, the, the Danish history, just to take a brief example,
- 1:27:21was really back to the, the, the Second World War and each
- 1:27:25country had their role in, in, in that.
- 1:27:27I know that Germany is very cautious about protecting
- 1:27:31personal data because they know that they can be misused and
- 1:27:35Denmark has a different approach to that.
- 1:27:37And I think that each country also because of local IT
- 1:27:41companies and, and, and, and that kind of of, of interests
- 1:27:47really effects our, our choice of, of systems.
- 1:27:51And it, it might not be wrong, but it is definitely worth
- 1:27:55giving a thought that we really don't give away what is valuable
- 1:28:00and what should be protected and, and, and we should have
- 1:28:04handled it otherwise. But I cannot say that one
- 1:28:06approach is right and another one is, is wrong.
- 1:28:08I think Mark, you have a good take on on this one as well.
- 1:28:14No, I just wanted to say in connection with what you said
- 1:28:17that in countries like Denmark, we're very trusting of our
- 1:28:23authorities because we haven't seen, we have seen a couple of
- 1:28:26cases of, but in general we have an understanding of ourselves of
- 1:28:32being totally free of corruption.
- 1:28:34This is of course an illusion, but, but, but this is what we
- 1:28:38think. And, and this also, I think it,
- 1:28:43it, it also feeds into our understanding of our
- 1:28:49relationship with the big tech companies.
- 1:28:51I mean, what are we giving them? We're thinking, OK, we can give
- 1:28:54them everything. And we have to a large extent
- 1:28:57already given them a lot of information about ourselves.
- 1:29:02And I think we need to be educated in Denmark.
- 1:29:04We don't necessarily because it is a balance, because there's a
- 1:29:07lot of good things about trusting authority as well.
- 1:29:11It it makes going through crisis and so on more, more easy.
- 1:29:17But but we do need to know like, and I think core you've been,
- 1:29:21you've been saying this like we need to understand what's going
- 1:29:23on here. What is the exchange we're
- 1:29:25making? OK.
- 1:29:29Just a strong opinion on this as well.
- 1:29:32I also have some questions for tours and guys, please so we can
- 1:29:36you know, we can do go on Yeah, thank you Tosun, this is for
- 1:29:42you. When AI challenge a student's
- 1:29:45relationship with uncertainty, what long term consequences do
- 1:29:49you foresee for their personal resilience and mental health?
- 1:29:57Tosun OK, he. This is can I have the question
- 1:30:08again? Trust me, half.
- 1:30:10Of it, yeah. Yeah.
- 1:30:11So it says like that. When AI challenge a student's
- 1:30:15relationship with uncertainty, what long term consequences do
- 1:30:20you foresee for their personal resilience and mental health?
- 1:30:27Well, the mental health, I can, I can tell you about this.
- 1:30:36They, they're going to be, I don't know if they're going to
- 1:30:40be or not, but I, I, I can see it that on this way, the about
- 1:30:46half of the student I think are using the AI And what is
- 1:30:53happening is because the, the, the AI is so perfect in in, in
- 1:31:01in the answer like like used to also talk about.
- 1:31:11It makes the student, it's going to be a, you know, lower self.
- 1:31:17What do you call it on English self?
- 1:31:21Esteem. Have to Yeah, yeah, yeah.
- 1:31:24And it's it's going to going to make you a second in, in in your
- 1:31:35mind that the first you have to to think about.
- 1:31:38I need to ask the machine before I do something because I don't
- 1:31:43know how and this know how in in in just asking the questions
- 1:31:52instead of being take care of or, or what do you call it to to
- 1:32:04stay in yourself. This is going to be a very
- 1:32:07difficult if it's going to be. So we first have to to ask our
- 1:32:14digital twins what we are going to do.
- 1:32:18Thank you, Charles, and I really appreciate it.
- 1:32:20I have a question also for Mark, please.
- 1:32:25OK. Yeah, it says you discussed with
- 1:32:27a Commission blindness. Are there systemic or
- 1:32:31architectural ways to measure if a student has lost the ability
- 1:32:35to recognize when they don't truly understand a concept?
- 1:32:42In a sense, this is, I think this is more question for for
- 1:32:44used who's dealing with this on a daily basis.
- 1:32:47But I think one point I'd like to make is that it's in a sense,
- 1:32:52it's not that different from what we, I mean, it's not a new
- 1:32:55problem. I mean, you've talked about this
- 1:32:58as well. We used to just copy the guy
- 1:33:00sitting next to you or we buy some essay from another student
- 1:33:06And, and, and the teachers and the professors have to be able
- 1:33:10to detect this along the way. And I for me, I mean you can,
- 1:33:18when you listen to used, you can, you can tell that there are
- 1:33:23structures in place at least with the individual teacher in
- 1:33:26form of experience and so on that allows you to detect
- 1:33:31certain patterns. But basically, but basically it
- 1:33:37is the same. I think like you says, it's very
- 1:33:40much in the perception that makes the difference.
- 1:33:43It's not much in, in the problem itself.
- 1:33:46The problem itself is the same as it's always been.
- 1:33:48But correct me if I'm wrong. Yeah.
- 1:33:53I mean, I mean, there's, there's a saying, Mark that, you know,
- 1:33:56nobody knows, there's not much about economics as a first year
- 1:34:00student of economics because they read one book, right And it
- 1:34:04made perfect sense and now they know it all.
- 1:34:08And I think sort of, you know, that that you, you get that
- 1:34:11effect with AI as well that you get 1 answer.
- 1:34:16Yeah. And, and you can definitely, you
- 1:34:18can definitely see sort of, you know, the different layers and
- 1:34:23interconnections. If they are lacking, you know,
- 1:34:26if the, if the relevant criticism is, is there, you
- 1:34:32know, that's an example of understanding.
- 1:34:34If the criticism that comes up of the model or the the own self
- 1:34:38critique of of the proposal is off, you know, then then you
- 1:34:45know, that's a, that's a clear tale in some ways that the
- 1:34:49understanding is not there. Whether that understanding is,
- 1:34:52is based off of AI usage or you know, something entirely
- 1:34:56unrelated to that, don't know, But that's, that's sort of one
- 1:35:00of the of the typical ways to to approach that topic.
- 1:35:05Thank you. Yes, I have another interesting
- 1:35:07question for Mark. How do we prevent the creation
- 1:35:11of a generation of leaders who have outsourced their critical
- 1:35:15judgement before they have even entered the professional
- 1:35:20workforce? Yeah, I think, I think we've
- 1:35:25already answered this is a lot of way because it it, it has a
- 1:35:28lot to do with education. If we let the students leave the
- 1:35:33educations without having the actual understanding, then we
- 1:35:36have a problem. And it will, it will, I think it
- 1:35:39will very quickly assert itself in reality when we start seeing
- 1:35:45people out there who doesn't understand what they're doing,
- 1:35:47who doesn't know how to deal with problems, who doesn't know
- 1:35:50how to assess their own abilities or other people's
- 1:35:54criticism and so on. So we will, I think we will, we
- 1:35:57will see it very, very fast if this goes wrong.
- 1:36:00And the basis of this is education.
- 1:36:03We need to make sure that the education is there.
- 1:36:06We need to make sure that that young people are educated and
- 1:36:10and keep on having their the trust and the abilities in their
- 1:36:15own faculties. That's basically it.
- 1:36:20Yeah. It's a, it's a good point.
- 1:36:24And the way I see it is that education does not only happen
- 1:36:28in schools. So that's also a perspective.
- 1:36:31I mean, I'm, I'm, I'm the father of three and, and, and they're
- 1:36:34adults by now, but I've seen that during the years they have
- 1:36:38learned so much from YouTube, they have learned so much from
- 1:36:41other sources. And, and, and it need not be a
- 1:36:44problem, but it is definitely a theme because once you get
- 1:36:49content that is not originally made for our teaching that might
- 1:36:54be based on different interests. Well, we should definitely also
- 1:36:59do our exercise in critical thinking.
- 1:37:02And it takes a lot of critical thinking to navigate in this
- 1:37:05world. It always did.
- 1:37:06And and it seems that the demand for critical thinking just keeps
- 1:37:11growing. And I so much hope that the
- 1:37:14schools will teach students critical thinking because once
- 1:37:18they manage to well, we can learn so much and we can
- 1:37:21navigate the world, which is complex and which really takes
- 1:37:25that. I'm quite sure that you are used
- 1:37:27also has a take on that, and probably also you and Mark.
- 1:37:33Yeah. And I think sort of, you know,
- 1:37:35that there's, there's an interesting side point here
- 1:37:40because I mean, critical sinking is, is one of those terms.
- 1:37:44It's thrown around the education sector.
- 1:37:45It's also one of those terms where when you go down and you
- 1:37:48try to define, you know, what is it actually?
- 1:37:52People struggle with a clear common definition.
- 1:37:55There are many good definitions out there.
- 1:37:58But but, but sort of. What is it actually?
- 1:38:02And and and. Sorry.
- 1:38:03About that I actually have in connection to this, I apologize
- 1:38:06used I have with these questions, which is truly
- 1:38:09connected to what you were just saying.
- 1:38:12Thank you. It says how do we ensure that
- 1:38:1421st century skills like creativity and critical thinking
- 1:38:18don't just become buzzwords, but remain qualities we can actually
- 1:38:22develop and measure in an AI driven environment.
- 1:38:26So yeah, here we go. And I mean, I think sort of, you
- 1:38:31know, the answer that was going there for Morella is, is that
- 1:38:34it's, it's going to be buddy, buddy difficult because I think
- 1:38:37AI tools are now in a, in a phase where they can even mimic
- 1:38:42critical thinking. So if I take Mark's example
- 1:38:44where he says, you know, include something, you know, about an
- 1:38:46objection the AI raised and you know, your response to it, well,
- 1:38:49I can prompt an AI to do that as well for me, right?
- 1:38:52There was a couple of years back when I, I thought sort of, you
- 1:38:55know, the easy solution to, to AI is simply to have people
- 1:38:58attach an addendum of the prompts that they used, right.
- 1:39:02But then I thought, you know, if I'm asking a machine to prompt
- 1:39:05my essay, why don't I ask it to prompt that addendum as well?
- 1:39:08And you know, sort of that idea went down the drain.
- 1:39:11So I think in short term, a lot of critical thinking, we need
- 1:39:17actually to go back to Socrates meetings.
- 1:39:19So we need to ditch the books. We need to go oral, we need to
- 1:39:21go old school. We need to have dialogues and
- 1:39:24and lively debates. Unfortunately, you know, that is
- 1:39:26happening in schools. But there's another point that
- 1:39:29I'd like to connect this with as well, because I think critical
- 1:39:33thinking, we talked about culture before different
- 1:39:35national approaches. You know, if we believe there's
- 1:39:37something that's culture, critical thinking might also be
- 1:39:40different in different cultural settings.
- 1:39:42Yeah. And I think one of the things
- 1:39:45that we are also missing right round now, and this is in
- 1:39:48relation to the point about sort of AI seems to coalesce the
- 1:39:52acceptable answers around a more narrow range.
- 1:39:56This narrow range is US heavy in its cultural exposure.
- 1:40:02Yeah. So that also means, you know,
- 1:40:05that as Rammstein was singing 20 years ago, we all live in
- 1:40:08America now. Well, in in the AI world, we
- 1:40:10definitely do. Yeah.
- 1:40:13And, and we don't notice it. I mean, there was a lot of
- 1:40:15writing when when DeepSeek came out.
- 1:40:18You know, the Chinese, they've made, you know, an AI and they
- 1:40:21made it cheaper and they made it faster than the Americans.
- 1:40:24Oh my God. Chinese people make something
- 1:40:26faster than cheaper, right? But but the point there was
- 1:40:29people were quick to point out and laughs, you know, see,
- 1:40:31there, there you have the marks of the Chinese censor there, you
- 1:40:34know, it has certain opinions about the Communist Party of
- 1:40:36China there. It has certain opinions about
- 1:40:38what happened on TMN Square in that particular year.
- 1:40:41According to Deep Sea, it was a very nice summer.
- 1:40:43And wouldn't you like to ask about something else?
- 1:40:45But The thing is, ChatGPT has a strong bias as well.
- 1:40:51And I challenge all of the listeners here to notice it
- 1:40:55because it's subtle. Yeah.
- 1:40:57And it's not obvious. And so if we want to use our own
- 1:41:00critical thinking, that would be where I would put my finger
- 1:41:02down. I think it's also I it's just a
- 1:41:08personal observation by using AI quite a lot.
- 1:41:13I think we should think upon one important factor that as you
- 1:41:20all, all of you might know, I mean my guess there is it will
- 1:41:27be really optimistic to and a good way to just create starting
- 1:41:32by creating a problem formulation.
- 1:41:34So it should be starting with an objective, a purpose and then
- 1:41:39try to work on around that problem formulation as an
- 1:41:44objective. So you're going through a
- 1:41:46process and then you go towards the end of it.
- 1:41:49I, I really make it like really short, but it's too long to
- 1:41:53explain it in that way. And it has to be also shown the
- 1:41:57process and on how to do so. And we are not here
- 1:42:01unfortunately today to do that, but just talking about it at
- 1:42:06least today. But we are going to arrive there
- 1:42:09as well. We are going to offer to our
- 1:42:11listeners and subscribers on Rewired to Rise also those kind
- 1:42:15of processes that they will be available on our channel on
- 1:42:20Rewired to Rise beyond the mindset, yes.
- 1:42:25We should really do sometime in the future, we could do an
- 1:42:28entire episode about the bias and these related topics because
- 1:42:32there is so much to that story. And in my experience, there are
- 1:42:37many mentioning the word bias but really not understanding the
- 1:42:42depth of it. I might not so.
- 1:42:45So I think we should dig deeper into that.
- 1:42:48Definitely you should do a monthly theme on that in.
- 1:42:55It's funny you should mention that because I'm actually we're
- 1:42:58coming up with something in the vein of that next month.
- 1:43:06OK, there is someone who would like to add something else,
- 1:43:10because we are kind of achieved at the end of this live session
- 1:43:15for today. For my part, I think it has been
- 1:43:19such an interesting opportunity. I mean, I was dealing with these
- 1:43:24themes before we even start planning this event.
- 1:43:28I had the opportunity to dig deep and I'm very thankful for
- 1:43:31for that. I learned a lot from from both
- 1:43:34the, the, my own research and presenting it.
- 1:43:38I learned a lot from from you guys and the, and the questions
- 1:43:41and answers put something on top.
- 1:43:43I hope that the dialogue and the questions will also continue in
- 1:43:47our other format so that there's so much to gain from this
- 1:43:51community. And thank you for for, for, for
- 1:43:54setting it up, Mirela. It's such a gift.
- 1:43:57You're welcome. It's my pleasure.
- 1:44:00OK then thank you to everyone, my guests, my speakers, and all
- 1:44:05of you, those who are behind the screen.
- 1:44:08Thank you so much for being here with us today and see you at the
- 1:44:12next live session.