GEF Madrid 2024: Conversation: K12 Education for the Age of AI

Summary

Charles Fadel, founder and COO of the Center for Curriculum Redesign, discusses the impact of AI in education, emphasizing the need to avoid anthropomorphizing AI and to focus on its current capabilities rather than speculative future developments. He highlights the evolution of AI from symbolic AI in the 1980s to today's large language models, emphasizing the importance of integrating AI responsibly in education. Fadel calls for a modernized curriculum that balances traditional and modern disciplines, fosters critical thinking, and prepares students for a versatile future. He also advocates for leveraging AI to enhance personalized learning and teacher support.

Takeaways

• 🧠 The speaker emphasizes the importance of not being overly anthropomorphic with AI, cautioning against trusting it too much simply because of its human-like appearance or interface.
• 🌐 Charles Fadel, the founder and COO of the Center for Curriculum Redesign, highlights the evolution of AI from the 1980s to present, noting the significant increase in computational power and data availability.
• 🔮 Fadel discusses the current phase of AI as an engineering phase rather than a scientific one, indicating that we are now focused on applying AI rather than discovering new theories.
• 💡 He warns of the dangers of anthropomorphization, suggesting that we are prone to trust technologies too easily, which can lead to unforeseen negative consequences.
• 🚀 The speaker mentions the importance of understanding the capabilities and limitations of AI, especially its current inability to achieve general intelligence or 'super intelligence', which is still a work of science fiction.
• 🔑 Fadel stresses the need to focus on the probable applications of AI rather than the possible ones, advocating for a pragmatic approach to leveraging AI's current capabilities.
• 🏫 He argues that education systems need to adapt to the digital revolution, modernizing the curriculum to include more relevant subjects like data science and computational thinking.
• 🛠️ The speaker calls for a reevaluation of what is taught in schools, suggesting that traditional subjects should be supplemented with modern disciplines to prepare students for the future.
• 🤖 Fadel suggests that AI can assist in personalized education by providing teacher support, allowing educators to focus more on local adaptations and individual student needs.
• 🎨 Creativity is discussed as a human trait that, while AI can assist in incremental innovation, the leaps in imagination that lead to groundbreaking ideas are still uniquely human.
• 🌟 The importance of motivation and a sense of purpose in students is underscored, with AI helping to identify and support projects that align with global challenges and personal interests.

Q & A

What is the main concern Charles Fadel addresses in the field of AI and education?

Charles Fadel's main concern is the potential for anthropomorphization of AI in education, where people might overly trust and believe in AI systems due to their sophisticated interfaces and capabilities, leading to a lack of critical thinking and reliance on technology.

What does Charles Fadel mean by 'overly anthropomorphic' in the context of AI?

Being 'overly anthropomorphic' refers to the tendency to attribute human characteristics to AI systems, which can lead to an overestimation of their capabilities and a misplaced trust in their outputs.

What is the significance of the transition from symbolic AI to neural networks in Charles Fadel's career?

The transition signifies the evolution of AI technology from limited processing power capable of handling only three layers of neurons to the current capabilities of handling thousands of layers, reflecting the significant advancements in AI and its applications.

Why does Charles Fadel emphasize the importance of not expecting general intelligence or super intelligence from AI systems soon?

Charles Fadel emphasizes this because he wants to highlight that current AI systems are still narrow and specialized, lacking the ability to generalize and reason across different domains, and thus should not be mistaken for having human-like general intelligence.

What is the 'engineering phase of AI' according to Charles Fadel?

The 'engineering phase of AI' refers to the current stage where AI is being developed and applied to solve specific problems, as opposed to the scientific phase which involves the foundational research and theories of AI.

How does Charles Fadel view the impact of technology like SMS and Twitter on society?

Charles Fadel views the impact of technologies like SMS and Twitter as having subtle but significant societal effects, such as changing communication habits and potentially undermining democracy, highlighting the need to be aware of the long-term consequences of technological advancements.

What is the 'zigzag boundary' Charles Fadel mentions in relation to AI capabilities?

The 'zigzag boundary' refers to the unpredictable and non-linear limitations of AI capabilities, where AI can perform exceptionally well in some areas while struggling in others, indicating the complexity and the need for careful consideration of AI applications.

Why does Charles Fadel argue that we should focus on the probable rather than the possible when discussing AI?

Charles Fadel argues for focusing on the probable to ensure a grounded and practical approach to AI development and application, avoiding the pitfalls of overhyping and misunderstanding the current capabilities and limitations of AI.

What does Charles Fadel suggest as a 'wise insurance policy' for education in the face of AI advancements?

Charles Fadel suggests that a wise insurance policy is to develop versatility in students, providing them with a broad range of capabilities and a strong base of knowledge and skills, enabling them to adapt to various situations and challenges presented by AI and the changing world.

How does Charles Fadel view the role of schools in a world with advancing AI?

Charles Fadel views schools as centers of stability, which are crucial for nurturing students' abilities and sense of purpose. He emphasizes that schools should not disappear but rather evolve to better prepare students for a world with AI.

What is Charles Fadel's perspective on the necessity of modernizing the disciplines taught in schools?

Charles Fadel believes that modernizing the disciplines is essential, advocating for the inclusion of modern subjects like data science, social sciences, and technology alongside traditional disciplines, to better prepare students for the future.

Outlines

00:00

🧠 Anthropomorphism and AI's Impact on Education

Charles Fadel, a global education thought leader, discusses the dangers of anthropomorphizing AI and the tendency to overly trust technology due to its human-like features. He emphasizes the importance of understanding AI's current capabilities and limitations, highlighting the shift from symbolic AI to neural networks with thousands of layers. Fadel also touches on the importance of addressing the real work of integrating AI in education, rather than getting lost in the hype cycle. He warns against the potential for AI to be misused or misunderstood, especially in the context of K-12 education.

05:01

🚀 The Evolution and Challenges of AI in Society

This paragraph delves into the evolution of AI, from its early days with limited processing power to the current era of large datasets and specialized processors. The speaker discusses the progress of AI from knowledge representation networks to large language models, noting the punctuated equilibrium of technological advancement. Fadel also addresses the subtle societal impacts of technology, such as SMS leading to Twitter and the potential for technology to undermine democracy. He stresses the importance of being aware of these impacts and the need for caution in how we integrate AI into our lives.

10:04

🤖 The Current State and Future of AI Capabilities

The speaker clarifies misconceptions about AI, stating that general intelligence or super intelligence is not imminent. He describes the current phase of AI as an engineering phase, characterized by the ability to process vast amounts of data and perform complex tasks. However, he also points out the limitations of AI, such as its need for data, tendency to hallucinate, and brittle reasoning capabilities. Fadel advocates for a balanced view of AI, one that leverages its capabilities while verifying its trustworthiness.

15:05

📚 The Role of Education in the Age of AI

Fadel discusses the importance of education in preparing students for a world with AI. He argues that schools are not going away and that they must adapt to provide a stable environment for students. Education must evolve to teach not just traditional subjects but also modern disciplines like data science and computational mathematics. The speaker emphasizes the need for a strong foundation in skills and character, as well as the development of a broad range of capabilities to prepare students for an unpredictable future.

20:05

🌐 Modernizing Education for the Digital Revolution

The speaker calls for a modernization of education to reflect the realities of the digital revolution. He questions why traditional subjects are still emphasized over modern disciplines like technology, engineering, and social sciences. Fadel suggests that education should be more about adaptability and versatility, preparing students to be Renaissance individuals with both breadth of knowledge and depth of expertise. He also discusses the need for professional development and assessments to make the teaching of 21st-century skills actionable.

25:08

🎨 Creativity and the Co-Creation with AI

Fadel explores the topic of creativity, suggesting that AI may be capable of incremental innovation but lacks the human ability for leaps in imagination. He uses the example of the progression from single-blade to multi-blade razors to illustrate how AI can excel at incremental analogies. However, he argues that AI cannot replace the human capacity for creative leaps, and education should focus on nurturing this ability in students. The speaker also touches on the importance of co-creating with AI, using it as a tool to enhance human creativity.

30:10

👨‍🏫 Personalized Education and Learning with AI

In the final paragraph, Fadel discusses the potential for AI to support personalized education, allowing teachers to focus on local adaptations and individual student needs. He mentions the development of chatbots trained on education research to assist teachers in the classroom. The speaker also talks about the importance of student motivation and the role of AI in helping students engage with real-world problems, such as the UN's Sustainable Development Goals. Fadel concludes by emphasizing the need for a nuanced understanding of AI's role in education and the importance of not anthropomorphizing it.

Keywords

💡Anthropomorphic

Anthropomorphic refers to the attribution of human traits, emotions, or intentions to non-human entities, such as AI. In the video, Charles Fadel warns against being overly anthropomorphic with AI, as it can lead to misplaced trust and reliance on technology that may not fully understand or share human values. He illustrates this by discussing how AI can have a 'nice face' and an 'unreal body,' which can lead people to trust it too much.

💡Artificial Intelligence (AI)

Artificial Intelligence, or AI, is the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. The video discusses the impact of AI on education, emphasizing its current capabilities and potential future developments. Charles Fadel mentions the evolution of AI from symbolic AI in the 1980s to the large language models of today, highlighting the significant progress and transformative potential of AI in various fields.

💡Hype Cycle

The Hype Cycle is a graphical representation of the maturity, adoption, and social application of a certain technology. In the context of the video, Fadel talks about the hype surrounding AI and how it often precedes a period of disillusionment before the technology's true value and applications are understood and adopted. He uses the Hype Cycle concept to discuss the current state of AI and to caution against overestimating its capabilities prematurely.

💡Knowledge Representation

Knowledge representation is the process of encoding information in a form that can be manipulated by an AI system. It was a significant aspect of AI during the 1980s, as mentioned by Fadel. The script refers to knowledge representation networks and symbolic AI, indicating the historical development of AI and its progression to more advanced forms of processing and understanding information.

💡Neural Networks

Neural networks are a set of algorithms modeled loosely after the human brain that are designed to recognize patterns. Fadel discusses the evolution of neural networks from being able to compute only three layers of neurons in 1989 to over 5,000 layers today, illustrating the exponential growth in AI's capability to process and analyze vast amounts of data.

💡General Intelligence

General intelligence, or AGI (Artificial General Intelligence), refers to the ability of an AI to understand, learn, and apply knowledge across a wide range of tasks at a level equal to or beyond that of a human. In the video, Fadel clarifies that AGI is currently a hype and not a reality, emphasizing that current AI systems are specialized and lack the broad understanding and adaptability of human intelligence.

💡Transfer Learning

Transfer learning is a machine learning method where a model developed for one task is reused as the starting point for a model on a second task. Fadel mentions transfer learning in the context of large language models, which can draw from vast datasets and apply knowledge across different domains, demonstrating a form of AI that can generalize and apply insights from one area to another.

💡Hallucination in AI

In the context of AI, 'hallucination' refers to the phenomenon where an AI system provides incorrect or fabricated information that it believes to be true. Fadel warns of this issue, noting that AI systems can sometimes cite non-existent papers or provide false information, which underscores the importance of verifying AI outputs rather than blindly trusting them.

💡Punctuated Equilibrium

Punctuated equilibrium is a concept from evolutionary biology, but in the video, it is used metaphorically to describe the uneven pace of progress in AI, with periods of rapid change followed by periods of stability. Fadel uses this term to discuss the historical development of AI, noting that significant advancements in processing power and data handling took decades to materialize.

💡Educational Technology

Educational technology refers to the use of both physical hardware and educational theory in the analysis, design, development, implementation, evaluation, and management of teaching and learning processes and resources. Fadel discusses the role of AI in education, emphasizing the need for schools to be centers of stability and to adapt to the changing landscape of technology, preparing students for a future where AI is increasingly prevalent.

💡21st Century Skills

21st century skills refer to the competencies that are believed to be most important for individuals to succeed in the modern world. Fadel critiques the superficial adoption of these skills in education, advocating for a deeper, more systematic approach to integrating them into curricula. He discusses the need for professional development and assessments to ensure these skills are effectively taught and learned.

Highlights

The importance of not being overly anthropomorphic when dealing with AI, as it can lead to misplaced trust.

AI's current state is in the engineering phase, not the scientific phase, emphasizing practical application over theoretical development.

The progress in AI has been punctuated by periods of hype followed by the settling of reality, highlighting the need for grounded expectations.

The potential dangers of anthropomorphization, drawing parallels to the unforeseen societal impacts of SMS and Twitter.

The human brain's tendency to take shortcuts and the risk of relying too heavily on AI, leading to reduced critical thinking.

AI's current capabilities in pattern recognition and processing, with examples like AlphaFold's protein folding breakthrough.

The distinction between narrow AI and the broader capabilities of language models, which can process diverse data types.

The limitations of AI, including its data hunger, hallucination tendencies, and brittle reasoning capabilities.

The hype cycle of AGI (Artificial General Intelligence) and the need for a reality check on its current and near-future capabilities.

The potential of AI to transform various sectors through specialized algorithms and corpora of data, emphasizing the need for adaptability.

The challenges of forecasting new job roles and the rapid adaptation of both employers and employees to technological changes.

The need for education systems to focus on developing versatile and adaptable skills in students, rather than just knowledge.

The importance of modernizing educational disciplines to include technology, engineering, and social sciences alongside traditional subjects.

The concept of '21st-century skills' and the need for professional development and assessments to make these skills actionable.

AI's role in incremental innovation and its potential to assist in the creative process, while acknowledging human superiority in leaps of imagination.

The necessity of co-creating with AI, focusing on how it can enhance human creativity without promoting laziness.

The significance of student motivation and personalized education, using AI to support teachers in adapting to individual student needs.

The potential for AI to support teachers with education research, providing real-time assistance in the classroom.

The vision of learning with machines, where AI can help with personalized learning and catching up for students who need additional support.

Transcripts

shaping the future keeping us

[Music]

[Music]

human Global for

Liv for

fatal let me now switch to English to

introduce him he is the founder and COO

of the center for curriculum redesign we

would love to hear more from his

expertise on the vision of the impact of

the AI in the education in general and

in particular in the K12 Charles fedal

is a global education thought leader an

author a f futurist we can say too an

inventor and he's also an expert for AI

at the OAC CD so praise Charles welcome

you can now take this state thank

you thank

you so well I'm a human intelligence um

I'm here to mention to you that one of

the things we have to face is being

overly

anthropomorphic when we deal with these

things because we tend to trust them too

much just because it has a nice face and

quite honestly an unreal body uh right

I'm sure women will say wait a

second um we we have the tendency to

overly believe some of the Technologies

we create so there's always a hype

moment where we're all excited and there

cool toys like this that show up and

then reality settles and then the real

work starts so we're going to be talking

about what the real work is is about um

I've been lucky enough in my career to

be involved with AI since okay I'm going

to disclose my age

1989 when as uh Mark was saying earlier

uh the whole world was going into

symbolic AI because we did not have the

processing power at the time my AI

company neural networks only had could

compute only three layers of neurons now

we can compute more than 5,000 layers

ERS we have massive databases we have

specialized processors and the world has

changed but we also have to keep in mind

that progress is always this punctuated

equilibria we had knowledge

representation networks and symbolic AI

in the 80s then in the two 2010s it took

30 years for the processing capability

to catch up to use neural networks

efficiently and then another decade to

get into large language

models again all due to large data sets

and now we're moving into this uh

situation of AI systems which going to

describe in a second what we're not

going to see anytime soon is general

intelligence or super intelligence and

I'll explain why in a moment but you

know we really need to worry about the

here and now and the enormous

capabilities of the here and now so this

is the engineering phase of of AI now

it's not the scientific phase of AI now

so I'm grateful that organizations are

paying attention to the um problems of

the future perfectly spot on I wish we

had done that with social media quite

honestly uh and we're still letting

social media run

crazy but when I was talking about the

dangers of anthropomorphization and so

on is because a lot of the problems we

have come from things we do not suspect

many years ago I was involved with SMS

you know for cell phones what was the

problem you know we had a little bit of

bandwidth between voice channels what

harm can happen with 140 characters sure

5 cents or charact or a SMS or whatever

so we deployed

SMS 10 years later your daughter cancels

a a a meeting with you at the last

minute because she can Via SMS thank you

at the supermarket and she doesn't show

up another 20 years and we have Twitter

and now we can undermine

democracy so these are the subtle

situations that emerge out technology

that we have to pay attention to our

brains are lazy by good evolutionary

design our brains are 20% of the energy

of the body consumption for 5% of the

mass so the brain is exquisitly tuned

to be as lazy as possible as frequently

as

possible and so that's why we judge

people on their faces very quickly we

judge how they're dressed

Etc our brain takes shortcuts and the

enormous danger that very few people are

talking about is how we're going to be

taking these shortcuts into believing

toys believing the hallucinations that

these sophisticated toys are going to

give

us how students will take the shortcut

to think less rather than more and

that's an enormous challenge ahead of us

because it's always easy to take the

easy way

out

okay so very quickly I'm going to

explain the differences between what had

happened 10 years ago and what's

happening now and why this is so

important we you know technology already

can do enormous computations and they

can really deal with spaces that are

incredibly big you know the three

complexity is 10 to the 535 so 10 with

535 zeros worth of

possibilities obviously no human can do

this here Alpha fold was able to figure

out how proteins fold which was

extremely timec consuming for humans so

when it comes to pattern processing

recognizing a pattern and applying it

modern AI is fantastic

IC but this was narrow because it was

only tuned to a game or a specific

problem to solve this was narrow machine

learning but what has happened with the

emergence of language models the

boundaries are much bigger because they

suck up enormous data

sets and second they can also suck up

different data types mathematics and

history not just go or chess or protein

so as a result they're capable of both

expertise and transfer they can look at

the situation and see an

analogy which a human may not be able to

see so they're incredibly capable

systems but they still have problems

they're hungry for data they we don't

know how they converge to a solution

they still

hallucinate very often they'll give you

they'll site a paper that doesn't exist

with all that do exist so it's very

sneaky it did that to me the first time

like I don't I recognize all these

authors but where are the papers very

sneaky and it's really brittle when it

comes to reasoning it does not

generalize a lot of different things and

it can generalize some very well and so

you have to be really careful about the

strange boundary about what he can and

cannot do it's not a simple boundary

it's a a zigzag all over the place so

it's up to us to figure out how to

harness these capabilities but keep in

mind these things we cannot trust it's

very much like uh Arms Control trust but

verify same should be applied here here

you go Mark trust that verify for for

AI so I was telling you that the

generalized intelligence and so on was

uh was overly overly hyped and this is a

gardner showing that AGI is reaching its

peak hype gen is at the peak hype and

they're going to go back down and we're

going to do a lot of work the same way

that autonomous vehicles are now in

reality mode sure they can drive on a

straight line in Oklahoma but good luck

driving at the pl delal in Paris at rush

hour

so reality check

so AGR is hype super intelligence is

still sci-fi and I'm not the only one

saying this even these

guys are saying this problem is we

confuse the possible with the probable

so I'm here as an engineer to focus on

the probable not on the possible there

are plenty of people who are looking at

the the possible and hedging against

that that's wonderful I'm here to

Leverage The Enormous capabilities we

have and they're enormous because we

have plenty of different llms we have a

lot of algorithms that are changing we

have Corp corpora of data that are being

verticalized like for education for

health care for finance Etc we have much

faster Hardware that's coming on that's

why uh Emily there was able to speak so

quickly there's a technology called Gro

that allows you to have almost realtime

uh responses

and then we can link a bunch of data

sets

together this is incredibly powerful you

don't have to wait for AGI to say wow

what are we going to do with all this

and this is what's evolving

fast so even if GPT 5 starts plateauing

which in my opinion it will because

again it's going through its S curve

it's going to start plate until we

change to a different type of algorithm

or we start merging algorithms

so don't expect gbt 5 to be massively

better even if Sam Altman says so

because he's a sales a

Salesman but this is already enormous

and we have to figure out what to do

with all of this so here's the problem

you know we're always

behind imagine a soccer player that's

constantly chasing chasing the ball

sorry football player American football

player that's constantly chasing the

ball and that's the situation where

we're all always get caught into and

that's why I'm delighted to have been

working already withk because there's an

opportunity for Independent Schools to

move ahead and show the rest what can be

done and not be at the lagered side of

traditional education systems public

systems of course we've all seen a lot

of different research coming from many

different Horizons this is from the IMF

and it talks about exposure to Ai and so

on but this research is not very precise

we really do not understand what jobs

are made of when we make these big uh

big pronouncements what we do is we

confuse tests with tasks with

jobs just because you can answer a test

like Pak or Pisa the oecd doesn't mean

that you can do the task just because

you're doing the task doesn't mean you

can do the aggregation of tasks with the

coordination and the adaptation that all

of these tasks

need so we have this uh overly

simplifying mindset which is called in

Psychology the done in Krueger effect

where what we don't understand very well

we

tralize for example every night I watch

brain surgery on YouTube for

fun you know we all have our fun and so

I met a surgeon just two weeks ago a

gastro enterologist ologist and I said

hey I can uh I can do brain surgery on

you now I know how to do it all you have

to do is you cut you you drill you open

you cut you suck you close you're

done that's how naive we are when we

think about a job in general and so in

reality whether it's a teacher or a

neurosurgeon it's a lot more complicated

to do the job than to do the task than

to do the test and so a lot of the

debate we've seen so far are very

superficial and that's what drives me

crazy is that people say oh AI cannot

create oh AI can creat oh I can do this

very superficial

conversations and that's why we

basically spend a lot of time in my team

going through and I'll show you the

results of this we also are terrible at

forecasting the emergence of a new jobs

terrible this was done by the world

economic Forum oops sorry

uh in 2017 looking at the jobs that were

created in the past 10 years that it did

not

forecast he was unable to imagine that

someone would ever want to be a YouTube

influencer really it's is that a job and

yes yes it is for our people not me or

an Uber driver you know these these

things were not around so we're terrible

at Imagining the new jobs

and we're also terrible at realizing how

fast people adapt this is both employers

and employees worldwide and these are

their main concerns and look they're all

talking about the same things they're

all trying to adapt furiously to this

new world so we underestimate the new

jobs we underestimate the adaptation we

overestimate what we don't understand

there we go so we have this sort of

situation and people start freaking

out so

there's no such thing as AGI takes away

all the job anytime soon and when we say

all the jobs that's really poorly

defined again the oversimplifications

that we see around the

world but that still means that we have

to deliver regardless even if even if

all the jobs were gone what are you

going to do with your 10-year-old are

you going to leave the 10year old at

home just trying to figure out the world

or are you going to send it send him or

her to a to a place where they can be

scaffolded with their

education where they have a safe place

where they can nurture their abilities

their sense of purpose so that

eventually they know how to work with AI

you're not just going to leave them at

home and at age 20 magically okay I'm

ready now no so education is not going

to disappear education is still going to

be here better done but still here all

the more importantly also as the world

becomes less and less stable School

schools are a center of

stability they have to be conserved as a

center of stability for the students

they maybe it may be the only thing that

they have that's permanent and stable in

their life so please let's not think

that schools will disappear and it also

means that we still have to prepare for

jobs partially at the high school level

and certainly yes at the University

level so that doesn't

disappear until of course AGI takes away

all the jobs I don't know 50 years from

now right now I'm not worried about

that and as Nas was saying earlier we

have you know she was quoting the

Industrial Revolution we are at the

digital Revolution

level and we invented Mass schooling at

the time and now what are we inventing

we're really sophisticating the

education we have this is not radical

this is ambitiously incremental again

not radical not disruptive but

ambitiously

incremental that means sure we have to

wonder about how do we teach better but

we also have to rethink should what

should we be teaching in the first

place if we have ai capable of helping

us on all sorts of things well what

should we be

teaching

okay well if you don't know what the

future holds what is a wise insurance

policy a wise insurance policy is to be

versatile so whatever the the whatever

is thrown your way you can always react

properly so you want to develop a broad

range of capabilities in the student

like a Swiss army knife and you can

always sharpen that that knife for a

given situation later but you have the

basics which means you're really a

Renaissance person you have breadth of

knowledge and you have depth of

expertise and you can add more expertise

like an M over time because you have a

strong base to begin

with so what does the strong base mean a

strong base mean paying attention to

modernizing the knowledge but also

paying attention to skills in character

meta learning all the quote unquote soft

skills that everybody talks about but no

one does anything in a systematic

way so we're talking about creativity

critical thinking Etc we're talking

about curiosity in ethics we're talking

about metacognition Etc

so when it comes to modernizing the

disciplines well why do we teach so much

trigonometry in mathematics and we don't

teach data science or we don't teach

discret and discrete and computational

Mathematics those are actually easier

mathematics

algebraically than a lot of algebra or

calculus why do we not do

that well because we're not used to

doing that but that's what should be

done

why don't we don't why don't we teach

world literature world history and I

don't mean world literature from a

Spanish or English perspective only I

mean literally Chinese Thai literature

Korean literature all the literatures of

the world why don't we teach even if we

teach performing art we teach acting

sometimes debate very rarely if ever

comedy or

improv so if the world requires you to

adapt quickly to situations don't you

want to be trained in improv that's how

you react fast someone asked a question

you know exactly how to react that's the

world we going to be living in and so

you have to pay attention to

adaptability where does that come from

from special sets of

disciplines same for visual arts okay

but you also need to add modern

disciplines because the old disciplines

go back to the Greeks the Trivium and

the quadri and the Middle Ages and so on

we have not left space for Technology

and

Engineering we have not left space for

social sciences except as options they

have to be

mandatory we also sometimes teach

entrepreneurship and business as options

it has to be

mandatory entrepreneurship is the job of

the

future social sciences if you need to

understand yourself and other

others for a better world why don't we

teach these things why is there only an

option again why so much of the

traditional disciplines and not the

modern ones and of course here with um

withk we we're developing modules for

clean tech and nanotech and we're also

developing modules for the modern

mathematics uh branches that I describe

so you're already doing something in

that respect and more next year

so now let's talk about competencies

when I wrote 21st century skills back in

2009 I was very happy to see that it

became a meme used around the world but

I became also very skeptical and

disenchanted because when we studied

this with Brookings we found out that

none of these jurisdictions

none was doing any professional

development or

assessments none everybody just sends

PDF to the teachers and say good luck to

you and everybody around the world has a

learner profile or a portrait of a

graduate or the goals of our school

sometimes with Latin even but it's not

it doesn't it doesn't hit it's not

actionable and the point here is to make

it actionable and so yeah yes we have

some some teachers some of the time

doing ad hoc things with the presence

and intent of these competencies but

we're going for Quality via professional

development and for evidence via

assessments that's where we need to get

to so I mentioned that we spent a lot of

time analyzing these things for every

single one of these competencies we have

a number of definitional subcompetencies

and for every single one of them we've

done an analysis of whether AI can be

good at Ambi pursuing ambitious Visions

despike risks can AI be leading with

initiative can AI engage with others etc

etc to decide whether or not AI was

going to be replacing or complimentary

to us so you see the level of analysis

that it takes rather than saying AI can

or AI cannot it's trivial we have to

really go into the

details and yes AI can be very fearless

and AI can be very tireless but that's

not the entire definition of courage and

resilience it takes a lot more than

that so you see that's the level of

precision that we need to get to in our

education systems where we analyze

things and we're precise about

definitions we're precise about goals of

what we want to

achieve let's talk about creativity for

a

second are is AI going to be more

creative or less creative than you

humans okay let's let's do a quick PA

who thinks that AI is going to be more

creative than

humans okay who thinks AI is going to be

less creative than humans that means

everybody else okay less creative than

humans what if I told you that 95% of

human Innovation was incremental and

that AI can do it I mean you have the

response right

here why because a lot of innovation in

the human domain is innovation by

analogy and by

extrapolation AI can analogize better

than humans actually and extrapolate

better

humans we are going to be filling a

database of all the patterns of the

world and then train the AI on how to

invent

next many of these inventions are very

incremental let me give you an example

there's an object that both men and

women use pretty much on a daily

basis the Raz are

blades

right razor blade

so someone comes up with two blades

right one blade to two blades someone

else comes up with three blades what do

you think AI is going to do analogy four

blades five blades six blades that's the

simple progression that you've seen even

Gillette and others do I think the

maximum is seven blades I don't think

there's room for your face

anymore seven blades you just go once

and you're done

but when it comes to the rotating the

pivoting blade that's different because

now it's no longer an analogy it's a

leap in

imagination and that leap in imagination

is what humans are still good at better

than

Ai and that's what needs to be

trained however even modart even Picasso

have a bunch of U mundane very average

Things That No One listens to or looks

at right mozar did a lot of stuff that

no one listens to but every once in a

while you come up with a flash of

Brilliance you cannot come up with these

flashes of Brilliance unless you go

through also the boring stuff the

incremental stuff so tough

question is AI more creative or less

creative is you should really we should

have said Charles as the wrong

question the question would have been

how are we going to co-create with AI

how can it help us create without making

us so lazy that we don't do the hard

part to come up with the flashes of

Brilliance so you see it's a complicated

answer every step of the

way oh by the way uh the reason why I

have so this is just one of my patents

as an example but this this is really

what I like it's a book called The Dot

Peter uh the Peter Reynolds the author

mentioned to me how he created that book

He's fell asleep with his hand like this

and on a piece of paper when he woke up

he there was a

DOT of

ink and so what I would have done I

would have just thrown it away Peter

being very creative he's like hm what

can I do with this and he thought of a

book describing the in Innovation

process for

children okay you see this is going a

little bit more than the normal it's

like a comedian comedians always do

things that we laugh about because we

recognize them it's just that we stopped

our thinking just before the funny part

and they make us discover the funny part

in our

thinking that's all all of

that so to I'm approaching the end we

have to pay attention to motivation of

students and in and indiv individualized

education personalized education that

means paying attention to the sense of

purpose of the

student and because AI has very limited

agency and no purpose at least unless we

surrender that agency but AI has an

identity the data sets the algorithms

the user interface already give it an

identity we think they're not but they

do they have an identity like this we've

given this an identity and you knowe

that I'm very careful of not calling it

her I'm calling it

it battle you have to battle

anthropomorphization every step of the

way

and so your purpose as a student whether

or not you have AGI is to do what you

love but also what you're good at what

you can be paid for unless you have

Universal basic income good luck with

that and what the world needs so that's

how you push the students to do a

project that has a purpose that can help

humankind and we have a database of 34

40 Projects on the UN sdgs the goal

Global challenges for Humanity the etc

etc so you can see how you can help the

student choose what they want to focus

on they want to focus on cyberspace or

blockchain or health issues or food or

whatever this is how you bring them to

help with a real world

problem lastly we're going to be

learning with the machines that means

that we can have teacher support via AI

so that teachers can focus a lot more on

local adaptations and personalized

learning that means that we're going to

have chatbot like the one we have

designed here which is trained on

education research because name one

teacher that can go to an education

school to a teacher college and remember

everything from p voty and everything

else and know exactly when to push it in

the classroom at the right moment and

have oh special needs oh Charles has a a

shortterm memory problem what do I do

what if we had the chatbot to

help and adaptive learning can help with

students catching up if they haven't

understood quickly

enough and that's another gain of AI

where the bloom to Sigma problem can be

helped we'll have to figure out what is

the amount of teacher

support going from teacher only to full

auto

depending on the

situations and that's it uh I don't have

time to go through this whole thing but

if you want to take a

picture please feel free to

download all right thank you very

[Applause]

much thank you very much Charles

thank you so much that was fascinating

you know he's traveling from the US um

to join us here today as well as our

next speaker I will introduce you um

right away before that H let's now take

a a step from

school we we have um seen this right now

um to universities we need to try to

understand how can we become from an AI

University and for that we have an

expert one of the most you know

interesting um experts on the world in

this specific um