The AI Education Revolution is Coming – or is it? | Dr. Philippa Hardman | TEDxSantaBarbaraSalon

Summary

Phil, an Affiliated Scholar at Cambridge University, discusses the potential of edtech to revolutionize learning. Despite over 30 years of research in learning science, he notes a disconnect between theory and practice, with technology often reinforcing traditional teaching methods rather than innovating. He highlights the importance of active, personalized learning, and the need for educators to embrace AI to enhance, not replace, human learning. Phil also touches on the 'three teams' approach to AI in education: avoidance, detection, and embracing AI, advocating for the latter as the future of education.

Takeaways

• 🎓 Phil, an Affiliated scholar at Cambridge University, is a learning scientist with a passion for edtech and its potential to transform learning experiences.
• 🔍 Learning science, with over 30 years of research, remains largely inaccessible behind paywalls and Ivory Towers, creating a disconnect with practical learning design.
• 📚 Phil advocates for the application of learning science research to bridge the gap between theory and the design of effective learning experiences.
• 🕵️‍♂️ The optimal conditions for human learning, as researched by educational psychologist Benjamin Bloom, suggest a shift from traditional teaching methods to more active, problem-based learning approaches.
• 📈 Bloom's findings indicate that a change in instructional methods could significantly improve learning outcomes, especially for non-traditional or underrepresented learners.
• 🚀 Despite the potential, educational technology has often reinforced traditional teaching methods rather than innovating new, more effective systems.
• 🤖 AI in education has been used to automate and accelerate existing, sometimes ineffective, practices rather than to innovate pedagogy.
• 🔎 Phil questions why, with the knowledge of effective learning formulas, technology isn't used more effectively to achieve these scenarios and provide equal learning opportunities.
• 🌐 The discussion highlights the need for a discourse between education technology and the application of AI to truly innovate and improve learning outcomes.
• 🔄 The 'three teams' concept discusses the different approaches to AI in education: avoiding AI, detecting and banning it, and embracing it as a tool for enhancing learning experiences.

Q & A

What is Phil's official title and what does it imply?

Phil's official title is 'learning scientist', which implies that he is involved in the study of how humans learn, a field that has been researched for over 30 years but remains largely unknown to the general public.

Why does Phil believe there is a lack of intersection between learning science research and educational practices?

Phil believes there is a lack of intersection because the research in learning science, which has been conducted for over 30 years, remains locked behind paywalls and Ivory Towers, not influencing the way learning experiences are designed.

What was Benjamin Bloom's significant finding about optimal learning conditions?

Benjamin Bloom found that the optimal learning experience for humans involves active learning, problem-based approaches, and ideally one-to-one coaching support, which significantly improves learning outcomes, especially for non-traditional or underrepresented learners.

How has educational technology historically been used, according to Phil?

Phil states that educational technology has historically been used to reinforce the traditional 'sage on the stage' teaching model, making it more efficient rather than introducing new, more effective systems of learning.

What is the risk Phil identifies with the use of AI in education?

The risk Phil identifies with AI in education is that it could make us more efficient at ineffective practices, rather than innovating pedagogy and improving learning outcomes.

What does Phil think about the potential of AI to disrupt education?

Phil believes AI has the potential to disrupt education and deliver on Bloom's vision of optimal learning, but the real question is whether we will allow it to do so, as we have had the capability to do this for 30 years and have not yet fully embraced it.

What is Phil's perspective on the recent integration of AI into learning experiences by Khan Academy?

Phil sees the integration of AI by Khan Academy as a positive step towards realizing Bloom's recommendations from 1984, offering personalized, adaptive learning paths powered by AI.

What are the 'three teams' Phil mentions in the context of generative AI in education?

Phil mentions 'three teams': Team Avoid, which tries to circumvent AI; Team Bennett, which attempts to detect and ban AI-generated work; and Team Embrace, which seeks to integrate AI into the classroom positively.

Why does Phil think Team Embrace is the most promising approach to AI in education?

Phil believes Team Embrace is the most promising because it acknowledges the inevitability of AI in education and focuses on educating students about AI, its uses, weaknesses, and risks, rather than trying to avoid or ban it.

What does Phil suggest is the role of educators in an AI-powered future?

Phil suggests that educators in an AI-powered future should play a critical role in guiding students on how to use AI, understanding its capabilities and limitations, and navigating its ethical and practical implications.

Outlines

00:00

🎓 Introduction to EdTech and Learning Science

Phil, an Affiliated Scholar at Cambridge University, introduces himself as a learning scientist with a keen interest in edtech. He discusses the disconnect between over 30 years of learning science research and its application in educational design. Phil's work aims to bridge this gap, focusing on creating technologies that facilitate the design of effective learning experiences, akin to those provided by the world's best professors or learning scientists. He shares a story from 1984 about educational psychologist Benjamin Bloom's research on optimal human learning conditions, which advocated for a shift from traditional teaching methods to more active, problem-based learning approaches, especially beneficial for underrepresented learners. Despite Bloom's findings, Phil notes that educational technology has largely reinforced the old system rather than introducing new, more effective methods.

05:02

🤖 AI in Education: Accelerating the Old or Innovating the New?

Phil critiques the use of AI in education, which has been around for 60 years but has mostly been used to automate and accelerate traditional, less effective teaching methods. He warns of the risk that AI could make these practices even more efficient without improving educational outcomes. Phil references Daniel Schwartz's view that the biggest risk of AI in education is its potential to reinforce ineffective practices. He contrasts this with historical examples like the printing press, which initially perpetuated existing ideas and power structures rather than fostering innovation. Phil argues that AI has the potential to disrupt education positively but questions whether we will allow it to do so, given our past tendencies to use technology for continuity rather than change.

10:02

📚 The Emergence of Team Embrace in EdTech

Phil discusses the three camps that emerged with the rise of generative AI in education: Team Avoid, Team Bennett, and Team Embrace. Team Avoid reacts to AI by avoiding its use, often through policy changes that restrict its application in education. Team Bennett engages in a cat-and-mouse game with AI, using it to detect and prevent plagiarism but facing challenges as students find ways to circumvent detection. Finally, Team Embrace, which Phil sees as the future, involves educators and institutions actively incorporating AI into the classroom to enhance learning experiences. He highlights the importance of educators' roles in guiding students on the ethical and effective use of AI, preparing them for a future where AI is ubiquitous.

15:04

🔍 The Future of AI in Education: Team Embrace's Vision

Phil emphasizes the growing prominence of Team Embrace, which advocates for the integration of AI in education. He references a study showing that AI, like the steam engine, has the potential to significantly increase human efficiency. Phil believes that AI will become an integral part of our work and lives, and that educators must play a critical role in teaching students about AI's capabilities, limitations, and risks. He suggests that the future of education will likely involve AI-powered learning, and that embracing this change is more sustainable and beneficial than avoiding or banning AI technologies.

Keywords

💡Edtech

Edtech refers to educational technology, which encompasses the use of both physical hardware and digital tools to enhance learning experiences and outcomes. In the video, the speaker discusses how edtech has been traditionally used to reinforce traditional teaching methods rather than to innovate and introduce new, more effective pedagogical approaches.

💡Learning Science

Learning Science is a field that investigates how humans learn and applies cognitive, psychological, and educational research to the design of educational interventions and technologies. The speaker identifies as a 'learning scientist' and emphasizes the importance of applying over 30 years of learning science research to the design of better learning experiences.

💡Ivory Tower

The term 'Ivory Tower' metaphorically refers to a place of intellectual pursuit that is disconnected from the practical matters of everyday life. In the context of the video, the speaker uses it to describe how research in learning science remains isolated within academic circles and is not effectively applied to real-world educational practices.

💡Bloom's Taxonomy

Bloom's Taxonomy is a classification of the different levels of cognitive complexity and learning objectives. The speaker references Benjamin Bloom's research on optimal conditions for human learning, which has significant implications for educational practices and the design of learning experiences.

💡Active Learning

Active Learning is a student-centered pedagogical approach that emphasizes active exploration, collaboration, and engagement in the learning process. The speaker discusses how Bloom's research indicated that active learning, as opposed to passive knowledge transfer, is more effective for all learners.

💡Personalized Learning

Personalized Learning is an educational approach that tailors instruction to individual students' needs, skills, and interests. The video highlights how technologies could be used to facilitate personalized learning paths, which is a key aspect of effective learning experiences according to the speaker.

💡AI in Education

AI in Education refers to the application of artificial intelligence technologies to enhance teaching and learning. The speaker expresses concern that AI has been used to automate and accelerate traditional, less effective educational practices rather than to innovate and improve learning outcomes.

💡Disruptive Innovation

Disruptive Innovation describes a process by which a product or service takes over a market by displacing established competitors with new technologies or business models. The speaker questions whether AI will finally be the force that disrupts traditional education models and fulfills the potential for improved learning experiences.

💡Generative AI

Generative AI refers to AI systems that can create new content, such as text, images, or music, based on existing data. The speaker discusses the recent rise of generative AI and its potential impact on education, including both the opportunities and challenges it presents.

💡Plagiarism Detection

Plagiarism Detection involves the use of software tools to identify instances where a work has been copied from another source without proper attribution. In the video, the speaker discusses how教育机构 have responded to the threat of AI-generated content by investing in plagiarism detection tools.

💡Team Embrace

Team Embrace refers to a group of educators and institutions that are choosing to accept and integrate AI technologies into their teaching practices rather than avoid or ban them. The speaker sees this approach as the future of education, where AI is used as a tool to enhance learning rather than as a threat to be controlled.

Highlights

Introduction of Phil, an Affiliated scholar at Cambridge University, with a focus on edtech and the intersection of learning theory and learning experience design.

Phil's official title is 'learning scientist', a field that has over 30 years of research but remains largely unknown due to being locked behind Ivory Towers and paywalls.

Discussion on the lack of application of learning science research in the design of learning experiences.

Phil's role as a researcher, educator, and edtech founder aiming to bridge the gap between learning theory and practice.

A short story about educational psychologist Benjamin Bloom's research in 1984 on optimal conditions for human learning.

Bloom's findings that a problem-based, active learning approach with increased support significantly improves learning outcomes, especially for underrepresented groups.

Critique of how educational technology has been used to reinforce traditional teaching methods rather than innovate.

Phil's concern that AI in education risks making ineffective practices more efficient rather than transforming pedagogy.

Historical perspective on the printing press, which was used to perpetuate existing ideas and power structures for 200 years before contributing to modernization.

Phil's motivation to use technology to achieve great learning experiences for all students, not just the privileged.

The potential of AI to disrupt education and the question of whether we will allow it to do so.

The importance of putting the 'ed' back into 'edtech' and focusing on education rather than just technology.

Examples of innovative uses of AI in education, such as Khan Academy's integration of AI into personalized learning experiences.

The emergence of 'Team Embrace', educators who are looking to integrate AI into the classroom in a positive and constructive way.

The ongoing cat-and-mouse game between 'Team Bennett', which tries to detect and ban AI-generated content, and students who find ways to circumvent these measures.

The potential of AI to increase human efficiency by a significant margin, as seen in the early studies on chat GPT.

The necessity for educators to educate students about AI, its uses, weaknesses, and risks in a future where AI is ubiquitous.

Transcripts

hi Mark and hi everyone like thanks so

much for inviting me I'm very excited to

be here so yeah I'm Phil

um as you say I'm an Affiliated scholar

at Cambridge University uh but most of

all I am a uh someone who's been very

interested in edtech for an awfully long

time now as you as you mentioned I am

incredibly old

um I am like my official title is

learning scientist if that means nothing

to you then you can be absolutely

forgiven uh because learning science is

really one of the world's best kept

secrets

um we have now over 30 years of research

into how humans learn

um and yet that research remains locked

behind Ivory Towers behind paywalls and

for lots of very interesting reasons

that perhaps we will get into tonight

um but that I love to talk about

um there is a lack of intersection

between that theory and the way that we

design learning experiences and so

that's really what is driving my work

and as you say I am both a researcher

and an educator myself

um but also an edtech founder looking to

kind of explore how we can build

technologies that somehow help to break

down those those barriers and that lack

of intersection and make it easier for

anybody to design a learning experience

like like the world's best Professor or

learning scientist

um and if I may I just like to tell a

short story at the beginning

um take everybody back to 1984. uh I was

about four years old everyone else was

probably not born yet because as I say

I'm very old uh but yeah educational

psychologist uh Benjamin Bloom at this

point in time was doing had been

researching for around about two years

uh this really interesting research

question which I've since picked up uh

which is which was what is the optimal

conditions for human learning

uh Bloom kind of observed that through a

combination of precedent uh tradition

practicality we have inherited a system

of teaching and learning which I'm sure

is all very like familiar to all of us

which is all about uh a sage on the

stage

uh the teacher is the conveyor of

information

the student is the person who absorbs

the information and regurgitates it and

or perhaps uh restructures it reframes

it but that is the fundamental basis of

The Learning Experience what Bloom found

and got very excited about was actually

that the optimal learning experience for

a human is very different

and what he found is that for all

Learners and for me doubly excitingly

the um the effect was very significant

on all learners but especially

significant for

um non-traditional I don't like that

term but excluded Learners

um

if we change from this stage on the

stage uh knowledge transfer system to a

system of instruction that is more about

um

proposing problems projects learning not

by being told something but by being

asked something learning by exploration

by research learning by getting things

wrong and then correcting them learning

through comparison and discussion that

problem-based Active Learning approach

and importantly also uh increase support

ideally one-to-one coach support

transformed learning outcomes for

everybody but as I say especially

um for um kind of underrepresented

groups

and so we have this really great moment

where we're all like amazing we've

cracked the code let's change how we

teach and learn Bloom's like okay kind

of mic drop let's go uh change the world

and what's really interesting to me is

that that didn't happen

and so we have had education technology

for you know this concept of Education

technology we've had now for 30 40 years

so since bloom

but that technology has consistently

been used not to introduce this new

system which we know works better but to

reinforce the old system so to make us

better and faster at ineffective

practice effectively so if you think

about I mean this is not an impressive

list of stuff but if you think about Ed

Tech as like for example the overhead

projector or the PowerPoint or the

interactive whiteboard all of these

things uh they don't help us to deliver

more active personalized and adaptive

learning they help us to

right like deliver lectures

and what we're seeing is

the same Trend with AI

so

ai's been around in education now for

like 60 years

uh as we all know it has uh with the

with the rise of open Ai and generative

AI

um over the last six months it's really

come on to everybody's radar and it's

given everybody access to it which means

that you know discussion of it is

unprecedented but what we've seen over

the last 60 years that we've been using

AI in education and particularly the

last six months is that we're doing the

same thing we are building AI technology

not to innovate our pedagogy and

innovate our impact on outcomes but to

accelerate and automate this broken

knowledge transfer process so we're

seeing AI tools uh you know I'm sure

you've all seen them all but like make

more content faster uh generate a quiz

from content all of these things

um and there's a huge risk here and I

think this Daniel Schwartz who is I

think the head of Education technology

at Stanford something like he's a

professor said recently and it really

resonated with me that the biggest risk

of AI in education is that it makes us

much more efficient at ineffective

practice

and so yeah I'm really interested to

think about why that is I'm also very

interested to

to put on people's radar the fact that I

think often we think of Technologies

like AI as inevitably Innovative and

disruptive but that isn't the case

um and as a historian I'm really

interested in things like uh the rise of

the the printing press in the 15th

century sometimes we celebrate that as

this great Innovative moment when things

really changed and we started to see

modernization whereas in fact the

research shows that for the first 200

years that we have the printing press it

was used to perpetuate existing ideas

existing behaviors existing systems of

power

so we're really like in this really

interesting situation and what really

motivates me in this space is to think

about if we know these formulas are

great what great learning looks like uh

why aren't we using technology better

to achieve those scenarios to give every

learner the same opportunities that you

know privileged people who get to go to

Cambridge and wherever else also get

um and when people ask me this question

like Phil is AI finally going to disrupt

education

I think my answer is always that

I I don't doubt for a second that AI is

able to disrupt education AI is

absolutely able right now

to deliver on Bloom's Vision on this

vision of where every student achieves

more faster

the question is more like the bigger

question I think the more interesting

question is will we allow it to do that

because we've been in a position where

we could do that now for maybe 30 years

and for whatever reason we've used

technology as a force for shoring things

up

as a force for continuity rather than a

force for change

um and so yeah that's that's why this AI

Revolution you know TBC is so

interesting to me are we going to see

Faster Horses

or are we going to you know see finally

uh the evolution of the car and and and

and with that

um the delivery on the promise that we

will actually you know serve every

single student as well as we can in the

classroom

you've you've dropped like 800 questions

in there and that was all perfect thank

you so much and I want to remind people

they can ask questions in the chat we'll

get to those later

um what's Bloom doing now by the way

um I don't actually know if Bloom is

still with us lots of people lots of

people since Bloom and me included have

continued that research and so now we

know conclusively In Bloom like was the

the forefather of this that

um a learning experience that is active

uh and personalized and individually

coached is chef's kiss

um I guess is still he's still around on

the scene but very much is like the

forefather of this movement

so when when you started by saying you

know all of this is kept in the Ivory

Tower

what needs to happen because it feels

like we're in this perfect opportunity

for people in edtech to reinvent the The

Learning Experience

um two weeks ago at Ted Sal Khan uh

founder of the Khan Academy showed how

they had integrated AI into the learning

experience for kids and I think it just

floored everybody we we showed that just

last Saturday uh that it is very

interactive very experience based very

question based and it's and it wasn't

just for the student it was also for the

teacher and so that looked like a best

practice of ways to go forward did you

are you aware of the work he's doing

yeah absolutely

like when I watched that Ted Talk Back

it's like yes it's like finally are we

is it happening because what um Sal Khan

described is exactly what Bloom

recommended in 1984. it's this perfect

individual coaching encouragement

personalized pathways through to a

shared outcome

um and so yeah and we're seeing other

examples too so I know that Microsoft

have released a Reading Coach which

again which scene so where Khan Academy

is um car meago I think it's called is

designed to be used at home as a kind of

pre-class or a post-class extension to

learning

um there are more and more tools

um appearing like the Microsoft tools

the Reading Coach to be used which is

designed to be used by teachers in the

classroom to deliver the same sort of

again we all have one goal but we sit

and we pursue that goal powered by AI

um and take very personalized adaptive

paths to it and get different sorts of

feedback which drive us through

um different paths to the same goal so

you're right in that there is you know

it there are examples of innovation out

there there are examples of AI being

used in ways which drives learner

outcomes but I think we're also in a

situation where and this has been the

case ever since edtech existed where

there is still a thinker lack of

discourse between education technology

I've said it before and I will say it

again I'm sure after this but like one

of my one of my uh kind of taglines is

that we need to put the ad back into Ed

Tech I think often we we end up building

technology uh to solve an immediate pain

hair on fire pain because that's the

thing that sells it's like I need to you

know design a lecture tomorrow so let's

build a tool that will do that in

seconds

um and so it's you know most edtech

Works in that way but yeah Khan Academy

and Microsoft doing great work to really

highlight how we can use AI to innovate

pedagogy rather than just delivery if

that makes sense it makes total sense

and and one of the things that uh uh I

appreciated about your writing because I

follow the things that you're saying

because you're you're very good at

putting information out there for those

of us that are interested in this uh

this field but you have this concept of

three teams which I have repeated uh

since I first was exposed to it explain

to our our listeners here what the three

teams are

yeah so uh and in fact I think it might

now be two but let me let me give you

the story so yeah in formal education so

this is global it's K-12 it's he it's

further education as we would call it

it's education uh when uh generative AI

kind of caught fire at the end of last

year and came into the world we tended

to have like three camps emerge

um so the first Camp was uh Team avoid

um and so we saw an initial reaction

where

um Educators and it's understandable and

I was included in this you know it was

like the immediate response is like

shoot if uh if everyone can now use chat

GPT to uh write their essays then the

systems are broken so what we'll do is

uh we'll avoid this risk by getting

everyone back into the room so we saw

for example in Australia very rapid

change to higher education policy that

said okay from now on you write essays

and you do exams in a room and we'll

watch you the kind of you know

panoptican Bentham uh approach all the

other institutions uh did things like uh

require students to create images or

videos or have oral aspects to the

to the examinations

um I mean it gives you a sense of how

rapidly things are changing because now

we can create images using AI as well so

that's broken

so there's that there's that group

um and then there was team Bennett which

is really interesting and this is

ongoing this is a very interesting game

of cat and mouse going on which I've

been watching since about December where

um you know I guess I guess in line with

existing protocol

um a lot of Institutions education

institutions said hey if uh Chachi BT

can write essays then that's plagiarism

so let's detect it and then

um basically ban it so make that uh

misconduct and so we've seen a lot of

Institutions investing in in very

rapidly built ironically using AI uh

tools which can detect whether or not

your essay has been written by tragedy

BT so the GBP zero various turn it in

whatever

um and what's really interesting is that

as quickly as these things have been

built over here in December

uh Revenge of the Nerds were on YouTube

uh explaining to all the students how

you can actually kind of clean your data

so that it's not detectable anymore

um and this is ongoing and so now we

have at first it was kind of hacks

through YouTube and now we have uh

companies being built uh which are

student facing and they say like hey

we can answer it we can we can write any

essay and no one will ever know that

chat GPT wrote it and so this game of

cat and mouse is going on there's also

some really important research that's

happened to show that

um chat gbt detection Technologies are

biased they're biased towards certain

types of language certain types of

Concepts Etc so I think the key message

is that perfect AI detection is not

possible and team bannett are now kind

of starting to realize that that they're

caught in this cat and mouse game I

think some institutions are still in the

process of buying the tech and they're

now like what we're going to do

and then and then the emerging Group

which is getting bigger

um I would say by the month particularly

over the last two months is team Embrace

and this team instead of trying to

circumnavigate it or ban it kind of say

okay well what would it look like if we

embraced chat GPT in the classroom what

does that look like for me as a teacher

and for my students on the ground and

that really I think is if we had a

beautiful graph now it would be helpful

but I think I see the first two camps as

really having a very short shelf life

because it's not sustainable

here we have um so this team Embrace I

think is where the future is going

I am

people may have seen it but there was a

really interesting and one of the first

controlled studies that came out about

chat GPT came out maybe two three weeks

ago and what it found is that chat GPT

increased uh human efficiency by 35 some

people are saying that's a massive

underestimation now to put that into

context steam in the in the Victorian

period increased our productivity by 25

so there is there is no doubt in my mind

that the most likely future is one if AI

remains low cost or no cost is one where

our work and our lives are AI powered

and so team Embrace uh or the team that

kind of acknowledge this and acknowledge

that in that world we as Educators have

a really critical role to play in

educating our students like about what

AI is and how we use it and what its

weaknesses are what the risks are

um and that's that's exactly what I've

been doing and all the other members of

Team Embrace have been doing

um and yeah I'm going to get a T-shirt

and send it to you that says team

Embrace on it thank you for for helping

us understand this very complex and very

topical

conversation

absolutely