November 4, 2024 • 30 mins
Are you curious about transforming your approach to education with AI? In this episode, we sit down with Darren Coxon, who shares his journey from teaching English literature to becoming a pioneer in artificial intelligence for schools. Darren's AI journey began with an encounter with ChatGPT, sparking the founding of Node Education—a venture dedicated to helping schools develop AI strategies. He offers firsthand insights into projects in Tunisia, where integrating AI into the curriculum has opened up new possibilities for students and teachers. This is more than just a conversation about technology; it's an urgent call for educators, school leaders, and parents to adapt and thrive in a rapidly evolving educational landscape.
Darren will present at GESS Dubai 2024 on November 12-14 at the Dubai World Trade Centre.
We delve into the ethical and legal issues schools face, from AI biases to updating assessment practices. Darren shares practical ways to integrate AI thoughtfully, using accessible platforms like Magic School and Notion to enhance learning experiences. Whether you're new to AI or looking to deepen your understanding, this episode is packed with insights to help you leverage AI's potential while maintaining educational integrity.
Connect with Darren here: https://node-edu.com
https://www.linkedin.com/in/darren-coxon/
Teach Middle East Magazine is the premier platform for educators and the entire education sector in the Middle East and beyond. Our vision is to equip educators with the materials and tools they need, to function optimally in and out of the classroom. We provide a space for educators to connect and find inspiration, resources, and forums to enhance their teaching techniques, methodologies, and personal development. We connect education suppliers and service providers to the people who make the buying decisions in schools.
Visit our website https://linktr.ee/teachmiddleeast.
Tweet us: https://twitter.com/teachmiddleeast
Follow us on Instagram: https://www.instagram.com/teachmiddleeast/.
Hosted by Leisa Grace Wilson
Connect with Leisa Grace:
Twitter: https://twitter.com/leisagrace
Linkedin: https://www.linkedin.com/in/leisagrace/
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:05):
You are listening to the Teach Middle East podcast
connecting, developing andempowering educators.
Speaker 2 (00:14):
Hey everyone, welcome back to the Teach Middle East
podcast, or welcome to the TeachMiddle East podcast.
If this is your first timelistening to us, where have you
been?
Join the 20,000 listeners thatwe have and get on board, get
regular.
Today I am pleased to haveDarren Coxon on the podcast with
me and we're going to geek outabout AI.
(00:35):
He said I hope I don'tdisappoint you, but I think he
won't, because I followed hiswork on LinkedIn for a long time
.
This is the beauty of LinkedInis that you can secretly stalk
people for years and they haveno idea about it.
So when guests reached out tome and said, would you like to
interview Darren Kotze, I waslike, well, I sure do, and so
(00:57):
welcome to the podcast, darren.
Speaker 3 (00:59):
Thank you, Lisa.
I hadn't realized I'd beenstalked by you for so long.
I feel quite honored.
Speaker 2 (01:04):
No, it's my pleasure.
So here we go.
Where did it all begin, Darren?
Take us back and give us theelevator CV rundown.
Speaker 3 (01:17):
Yeah.
So I mean, like a lot of people, I wasn't expecting AI to come
and slap me in the face like itdid a couple of years ago.
None of us were, were we?
So I mean, look my background'seducation, 100% Teacher,
english literature actuallyNothing, techie, nothing.
Computer science.
I taught English and drama andmedia and film, et cetera, et
cetera, for many years, theoryof knowledge, and then I moved
(01:38):
into headship and then I gotinto managing groups of schools.
So I managed the BrightonCollege in Abu Dhabi and a few
others the academic side, notthe operational side.
My last role was managing, aschief operating officer, a group
of international schools.
We were based in Bahrain, butone of the schools was in
Tunisia and I was seconded thereto be executive head.
(01:59):
And I actually arrived inTunisia at almost exactly the
same time that CHAT GPT arrived.
So I came October November twoyears ago, and ChatGPT arrived
at the end of November of 22.
And I remember playing aroundwith it over the Christmas
holidays and thinking there'sjust something in this.
This is not interactivewhiteboards or VLEs or even
iPads.
(02:19):
There was a lot more.
I could see the potential todisrupt and, sure enough, within
a couple of weeks of itsrelease, new York State had
banned it and I thought, yeah,well, obviously people are
nervous about this.
So I decided, like I could I ama geek, even though I'm an
English teacher, definitely onthe geek spectrum just to go all
in and and what we did is we asa school, we we really went for
(02:40):
it with AI, so a lot of theteachers were using it really
well.
We then collapsed the curriculumat the end of the second term,
so June of 2023, people werelearning about sustainable
development goals with AI andthey were running, with passion,
projects with AI and it wasjust really successful.
I then decided to come back tothe UK and basically just do
what I'm doing now, which iswhat I learned over that almost
(03:03):
a year in Tunisia and what I'velearned since Just helping
people understand it, use it,not be scared of it, have a bit
of fun with it, don't take ittoo seriously it takes me so
seriously in education but thenalso be aware of the risks and
understand what the risks are.
Someone said to me it's likeit's all fear and fun with you,
derna, which I suppose is kindof.
Maybe that's my catchphrasefear and fun.
(03:24):
But yeah, definitely to makesure people are mindful of it,
but don't take it too seriously.
Speaker 2 (03:30):
So tell me then, what are you currently doing as it
relates to AI, because I knowrecently you went full time into
this and you launched something.
I told you I stalk you, that'sjust how it is.
So, because the listenersaren't up to date as I am.
What have you launched?
What are you doing now?
Speaker 3 (03:48):
Yeah, so I mean I've been doing the consultancy for
about a year now, on and off,more on and off.
So I came back to the UK sortof September, october of last
year, and since then I've beenreally been doing my own thing,
supporting some platforms, butmainly just doing my own thing.
So I've just I I've actuallyjust launched node, node
education.
So node is basically a supportfunction for schools and
(04:11):
colleges around strategy,because that's key the training,
staff training policy procedure, the kind of paperwork that
sits in the background, and thensomething around the tools as
well.
I'm actually developing some aitools at the moment.
So node is kind of a catch-allwhich which really supports in a
360 degree way, I suppose,schools and what on earth to do
(04:34):
with AI.
So that's kind of my way in.
But yeah, look, I, I'm speakingat guests, I speak all over the
world and I'm just here forschools to help them implement
it in a way that actually worksfor them.
Speaker 2 (04:46):
So we're going to go into the nitty gritties, as
school AI is now disrupting theway the students learn.
It's disrupting the way theadministrators do their work.
It's disrupting how teachers dotheir work.
What are some of the thingsthat schools need to become very
cognizant of around the use ofAI or an implementation of an AI
(05:11):
policy?
Give us the one, two, three ofit all.
Speaker 3 (05:15):
Yes.
Well, the first thing is thatnothing will change unless
people realise that there has tobe change.
So one of the first things aleader has to do is to create
this sense of urgency, and weknow this.
I mean John Kotter has beenspeaking about this for years
and years and years.
I mean his Leading Change bookfor me is one of the best books
on leadership ever written.
So one of the first things thatleaders have to do is make
(05:35):
staff understand that andparents and students understand
that this is this is thedirection of travel.
Love it or hate it, you reallydo need to be going in this
direction.
I think from there you need tomake sure you're well.
Number one is that the head,the head teacher, the principal,
has to be not necessarilytechnically leading this, but
has to be brought in.
I say to the schools that Iwork with the first thing that
(05:58):
needs to happen is the headneeds to stand up in front of
the staff and say this is thedirection of travel.
We really have no choice,because if you don't do that,
what tends to happen is you getlike pockets of use in the
school.
You'll get.
You get one or two teachers whoare using it kind of slightly
furtively, like they're feelinga bit guilty, because teachers
by nature, we all have a bit ofa guilty conscience.
We just do.
It's just part of who we are.
(06:19):
And so often you get teacherswho are using it in the
background and they're thinkingoh, this is working really,
really well.
And I know this because so manytimes when I speak at
conferences, I get people comingup to me and saying I've been
using it for this and it's beenreally, really effective.
And I'm like so what are youdoing as a school?
They're like well, nothingreally.
This is just me.
So I think you need to makesure that you've got that
(06:39):
urgency, that person at thefront driving it.
And then it's about who aregoing to be your implementation
team.
And one of the pieces of adviceI would strongly give to schools
is if you just create, I don'tknow, like an AI working party
and the brief of the AI workingparty is understanding how best
(06:59):
to implement AI across theschool it's such a vague and
nothing thing.
It doesn't mean anything.
Well, where, how, why?
My suggestion will be plan aseries of maybe it's a collapsed
curriculum day, or even just ahalf a day or some sort of
festival or something that thatworking party is working towards
.
So, rather than it being we'recreating a group of people to
(07:22):
bring AI into the school, we'recreating a group of people to
plan for an event where ai isfront and center.
It worked so well, unbelievablywell, in tunisia, so I would
strongly encourage that to bethe case.
But the other thing I would sayis that, when it comes to
actually bringing it into theclassroom, don't straight away,
because what will probablyhappen is it will go wrong
(07:43):
because it does, because ai isnot that good right now.
I mean does wait.
I mean it drives me mad.
I mean, for someone who workswith it all day, I'm just
forever.
I'm surprised I've got hairstill forever pulling my hair
out.
It's just, it does stupidthings all the time.
You don't want to do stupidthings in front of the kids, in
front of the students.
So I would say, allow your stafftime to play with it and just
(08:04):
see what it's good for.
I mean, I know what it's goodfor within my sphere, but I'm
not a mathematician and I'm nota physics teacher, so teachers
have to be given the time toplay, and so I think, after
establishing the sense ofurgency and making sure you've
got the right people planningtowards an event.
You've got to find the time.
People planning towards anevent, you've got to find the
(08:26):
time.
So one of my talks just to kindof skip forward to some of the
things I'm doing at Guess in afew weeks' time is basically
about how do you make room forAI as a school leader.
How do you find the time to giveyour staff the freedom and the
permission and the time, thephysical time, to actually play
with it, because if you don't,nothing will happen.
It will literally be one or twokeen, probably 30-something Gen
(08:50):
Z.
Generally it tends to be newmembers of staff who I really
like going for it.
But then you get the kind ofthe slightly older, basically my
age, sort of Gen X.
You get one or two that areinterested, but often you'll get
them like oh, I just don't havetime for that, I don't have
time for church.
You have to make time.
You have to make time becauseit's getting so much better and
(09:11):
it's so much more useful and somuch more challenging that if
you don't, what you're going tofind is your students are going
to use it and they're just goingto accelerate away from you and
you're just going to be leftbehind.
Speaker 2 (09:21):
I tell you what I've noticed.
I've noticed because I'm likealways on LinkedIn or X and
stuff.
I've noticed that there are,now, unknown to the school
leaders, groups of staff who'veput themselves in these online
groups and online chats who aredoing different things with AI,
all to the ignorance of theschool leaders, and they're
(09:44):
implementing and maybe havingtheir own ways of dealing with
it, without the school havingany knowledge at all.
Speaker 3 (09:53):
Yeah, that's absolutely true.
And what will then happen isthe teachers will either leave
that school because they're like, do you know what?
I've reached a plateau and Ican't get any further Well, they
just go off and do itthemselves.
And I've seen two or three muchyounger than me I mean I've
been around.
I mean this is year 26 for mein teaching and education.
They've only been in the rolemaybe five, six, seven years.
(10:13):
They've already becomebasically gone up and do
consultants because theyrecognize there's a real need
for people who have beenpractitioners, who've actually
been in the classroom, so not edtech salespeople, but actual
teachers who can then speak thelanguage of teaching and have
the credibility amongst theirpeers who then go off and
support other schools.
So you're absolutely right.
(10:33):
I mean I posted two days ago,two or three days ago, on
LinkedIn.
This is 100% on school leaders.
Anytime you get a large seismicpotential risk and I think the
last time we had it was exactly10 years ago when we had Keeping
Children Safe in Education wasreleased.
The first edition of that wasreleased in 2014.
(10:53):
Inspections then suddenly wentup a notch and that was
absolutely on school leaders tosort out to deal with, and
governors.
By leadership, I actually alsomean governance Right now, as of
literally, and governors.
By leadership, I actually alsomean governance Right now, as of
literally, and even from lastweek, when we saw the US parents
suing their son's high schoolfor accusing him of cheating and
(11:14):
it's allegedly apparently wedon't know the whole facts
caused him irreparable damage.
So they sued the school.
From what I hear, they may havewon, or certainly they're
likely to win.
I think, with all of thesethings coming and bubbling up,
we're at that similar inflectionpoint.
We're at an inflection pointthrough the 2010s and then 2014
KCSIE.
We're at a similar point 10years later, where they are only
(11:35):
, I think, much more potentiallydisruptive.
So heads and governors have totake this seriously.
For me, it's our number oneagenda item right now, above
anything else.
Speaker 2 (11:46):
I agree with you 100% .
And then we want to look atwhat are some of these risks
that heads need to be cognizantof, aware of.
Speaker 3 (11:57):
Yeah, so I mean the first one is exactly to the
point of the family who's suingthe school.
So I've been speaking on thisfor about a year and a half and
literally on this.
In fact, one of my slides formy presentation which is one of
the only slides I haven'tchanged but virtually every time
I present I present somethingdifferent because it's changing
all the time was that if you usean AI plagiarism detector, best
(12:21):
case and you confidently sayyou cheated and the student's
like, no, I didn't, no, Iabsolutely did not Best case
scenario you've destroyed thetrust between you and them.
And not only have you destroyedthe trust with that individual,
you've probably lost the wholeclass Because they all think,
well, we're just waiting for himto accuse us.
Now Worst case scenario is theparents sue the school, which is
(12:44):
exactly what we saw last week.
So I think one of the biggestrisks that we have is not having
any policy in place, not havingany staff training in place,
allowing staff to use plagiarismdetectors which do not work.
We know this.
We've been banging on aboutthis for months and months and
months and months.
These do not work.
Last week, I put the USDeclaration of Independence
through an AI plagiarismdetector, bearing in mind that's
(13:07):
like 1774, 1776 or something250 years old 50% generated by
AI.
I mean, you know the USDeclaration of Independence,
apparently, according to one ofthe GPT-00, I can't even
remember which one it was,doesn't matter 50% generated by
AI.
So I think that's probably thebiggest risk.
You cannot use them.
(13:27):
You cannot accuse students ofusing AI in that way, because
you will absolutely destroy thetrust amongst all of the staff
and the students and the parents.
You literally will.
You may as well just kind ofgive up.
So if that's the case, what doyou need to do?
Well, you need to think of,first of all, other ways to
assess.
So, ok, let's start to havethis conversation about other
ways.
(13:47):
So you may know Russell Cayley.
I think you may have eveninterviewed him in the past
Think Learning Studios.
He's done a lot of work aroundsort of portfolio.
So how do we start getting ourstudents to generate portfolios
of what they know, portfolios ofwhat they've done, portfolios
of what they understand, andthen start having conversations
around those portfolios andreflecting which is way more?
(14:10):
It's more of a real world thing.
Anyway, we're going to need tomove away from exams.
It won't be long before youknow what I'm wearing here.
I could easily have theaugmented reality.
I mean right now.
Those glasses are very clunky.
Give it five years, they won'tbe.
I mean even, maybe even lessthan five years.
So contact lenses we're goingto see these things happening in
(14:31):
the next three to five years.
So what we've just started tosee, literally as of this term,
is the beginning of the singlebiggest inflection point that
we've ever experienced ineducation.
And bear in mind we're talkingabout a system which is 250
years old.
This is where it has to change,because if it doesn't, we are
going to see more schools beingsued.
We're going to see higher ratesof drop, doesn't?
(14:51):
We are going to see moreschools being sued.
We're going to see higher ratesof dropout.
We're going to see universitiesrumbling, which they already
are.
We're seeing these thingshappening in real time.
So this is why I do what I do.
Look I, I, as I said, I'm anenglish teacher.
I never set out to be an aikind of consultant, a trainer,
but I can see the risk, so Ifelt like it was actually.
(15:12):
I always felt like I had atrainer, but I can see the risks
, so I felt like it was actually.
I always felt like I had amoral obligation to be part of
the solution, rather than kindof complaining from the
sidelines.
Speaker 2 (15:20):
Yeah, and talk to me about the biases of AI, because,
yes, there are the risks offalse accusations, of using it
when they haven't, and the AIdetectors are diabolical they're
terrible, they don't work.
But then what about the biasesthat come into AI that students
and teachers, and especiallyschool leaders, need to be aware
(15:42):
of?
Speaker 3 (15:43):
Well, I think I'm probably in the minority when I
would say that most generativeAI is actually becoming less
biased.
The reason is that they arethat the computer scientists are
doing everything they possiblycan because people see it as
being such a big-ticket item.
They're actually doingeverything they can.
So if you work with chat GPT,okay, I'm not getting a strong
(16:05):
sense of bias coming throughanymore, particularly when we
think about how we use chat GPT.
Okay, it's very transactional.
You want it to do something andit does it.
You tend not to go back andforth in the way that you might
with a sort of more personablechatbot.
So I would say that when itcomes to the likes of chat GPT,
(16:26):
I'm less worried.
Okay, it's not something that Italk much about.
It's not something that I thinkyou need to be too worried
about right now, and I do thinkit's not something that I talk
much about.
It's not something that I thinkyou need to be too worried
about right now, and I do thinkit's getting better.
Text to image AI yes, it stillhas some biases, but I would say
that actually they push thingstoo far the other way.
So, if you text, I don't knowif you prompt an app like a
platform like Adobe Firefly.
(16:47):
Often you'll get images whichare almost trying too hard to be
broadly represented, to thepoint where they're no longer
useful because you want it togive you something and it won't
because it will actually giveyou something, which is it tries
to be almost too unbiased.
I would say that where thebiggest concerns are with AI
more broadly we're not talkingabout generative AI more broadly
(17:08):
are the biases inherent in thealgorithms that we see in TikTok
.
I am, more broadly are thebiases inherent in the
algorithms that we see in TikTok, that we see in Instagram, that
we see in YouTube, the shortsthat are constantly pushing a
very narrow worldview on ourchildren.
Ok, so TikTok's the mostobvious example.
The biases, the way thealgorithm works.
(17:31):
Essentially, you know it worksout what you like and it just
gives you more of the same.
That that's the thing thatworries me.
So I almost feel like we don'ttalk enough about that.
We talk a lot about ai bias,but when we talk about it, we I
personally believe the damagingai bias sits more in the kind of
the machine learning end, inthe kind of the the algorithmic
end, rather than worrying thatchat gpt is going to give me
some sort of biased output Again.
(17:53):
I think I'm in a minority there, but certainly from my
perspective that's where I thinkeducators and leaders and
parents should be more worriedabout than chat GPT.
Speaker 2 (18:04):
You touched on something that concerns me
because, as a parent, I worryabout my kids' use of things
like TikTok and YouTube Shorts,etc.
I wanted to ask you, then whatshould parents be aware of in
terms of their child's use ofthese platforms?
Yes, you talk about themserving more of the same, but
(18:26):
what else should they be lookingout for?
Speaker 3 (18:29):
It's hard, isn't it?
Because if you allow your childto have a phone I mean, I have
a 17 year old, an eight-year-oldand a five-year-old.
So I've got one who's kind ofgot through that, she's almost
got through the worst.
She's pretty sensible.
But I've got an eight-year-oldboy.
He's very, very susceptible andprone to like boys are.
He loves minecraft and he'slearning to code and he likes
(18:51):
that whole online world.
And then my five-year-old she'sstill playing with dolls.
So I don't know which wayshe'll go, but it worries me.
But if, if you're a parent who'sgiving you a phone I'm not
judging either way, but lots ofchildren have phones then it's
actually very, very hard.
Because if you're allowing yourchild a phone, the only thing
you can do is maybe prescribethe times that they use the
(19:12):
phone.
But then, of course, that canbe a big challenge, because the
phone becomes an extension oftheir hand and how do you take
it away from them?
And I'm not here to give adviceon parenting, because that's,
quite frankly, it's not not myplace to, but for me personally,
the only thing I can think todo is to try and push as many
interests my children's way thatget them away from their phone.
(19:35):
I hope that one of them sticksand hope that carries them
through those very difficultyears when actually all they
want to do is just be absorbedon an online world.
So my 17 year old daughter shejust got buried into football
age eight, nine, ten years old,and actually that carried her
through her early teens becauseall she was doing she was
playing football.
She played for a local side,she was training with one of the
(19:57):
big UK teams on their juniorsquad.
That carried her through.
So the only advice I can giveis what I'm going to do myself
is try and get my childreninvolved in things that interest
them, which take them away froma phone, because once they're
in the phone, then what can youdo?
Speaker 2 (20:14):
Yeah, that's true.
For the record, my kids don'thave phones because no.
Speaker 3 (20:21):
Because, yeah, because exactly.
Speaker 2 (20:23):
Because no?
So taking it back to schoolsand taking it back to teachers.
There's a lot of fear around AIand some people have completely
buried their heads in the sandwith it.
But when it comes on to usingit in the classroom you already
said, not with the students, notin front of the students.
Speaker 3 (20:45):
Initially.
Initially, yes.
When you're first understandinghow to speak to it, do it on
your own.
Don't learn in front of thestudents.
It's not good for them yeah.
Speaker 2 (20:54):
Yeah, I agree, I actually agree with you.
But where should staff?
Let's say they're schoolleadership, they're yet to come
on board.
They're going to come on boardbut they're yet to do it.
Where should staff begin?
Because they want to getknowledge.
They want get knowledge, theywant to know how to utilize
these tools and they don't.
Speaker 3 (21:14):
They don't want to be afraid anymore.
I've slightly revised myopinion for maybe a few six
months ago, because six monthsago I was very anti-wrappers.
The wrapper app is where youhave ai in the background and
then you have some sort of userinterface nice friendly user
interface on the top, and youpush a button rather than doing
a prompt.
I would actually say now thatthey're possibly a good way in
when you're first looking tobring it into the classroom.
So the obvious one is MagicSchool.
(21:36):
Okay, so I think that some Idon't necessarily think all of
Magic School is fantastic, but Ithink there are some elements
which I really like.
So, for example, you can set upa little student room.
You can put AI tools insidethat student room and then you
can let the students use them.
So text-to-image generators,safe little chatbots.
So there are ways that you canuse platforms like Magic School
(21:59):
in quite a good way.
Gamma's another one.
So Gamma for presentations,that's fantastic.
Notion for documents, that'sreally, really good.
It's all got AI built in andit's done in quite a safe way.
Quizzes, which I think is thebest quiz.
I'm not affiliated to any ofthese, by the way.
So these are just ones that Iuse myself and that I recommend
that schools look at.
Quizzes is great for littleAI-generated quizzes, obviously
(22:22):
because it's called quizzes.
These are good, safe, simple,quite boundaried ways to bring
AI in AI to generate amultiple-cho AI in AI to
generate a multiple choice quiz,ai to create a presentation, ai
to share a little text, toimage generator with your
students, and they're cross ageas well.
So I mean you can use thosewith younger students as well as
(22:42):
older students, and you need tocheck in the terms and
conditions on the site, butthey're doing everything that
they can to make sure that theseAIs are nice and safe.
So I would say that's probablythe first thing I'd do.
Just try Magic School.
It's free.
The schools, they're all free.
By the way, I mean, unlessyou're going to go heavily into
AI proper, like I do a lot ofcoding with AI, so I code my
(23:09):
website from the base,completely coded with AI.
If you look at my website nodedash I can't remember what it is
node-educom that's allgenerated with ai.
Um, because I can't code.
But if you're not going to gothat far, then actually most of
the wrappers are going to suityou.
Well then, of course, onceyou've got the hang of it and
you realize the limitations ofthese wrappers, then you can
start to use chat gbt, claude,google gemini, a google ai
(23:30):
studio, because actually you'vestart to use ChatGPT, claude,
google Gemini, google AI Studio,because actually you've started
to get an understanding as towhat these wrappers are good for
and, quite frankly, whatthey're not very good for,
because there are somelimitations Bringing it into the
classroom.
So there are some really simplethings that you can do and I'm
dropping all of this onto mywebsite soon, all for free, and
(23:53):
I'm dropping all of this onto mywebsite soon, all for free
things like just setting up alittle assessment bot or a
little feedback bot for students.
So, rather than you marking thefirst draft of their essay, you
give them a prompt.
They put the first draft of theessay in with the mark scheme
and it doesn't give them anaccurate grade because it's
really not very good at grading.
So I would also suggest teachersdon't expect AI to grade papers
(24:15):
very well, because it won't,but it will give quite good
formative feedback.
So that kind of kills two birdswith one stone, because it
saves you time as the teacher,but it also slightly empowers
the student because suddenlythey're like they're working
with an AI and it's giving themideas and they're chatting with
it.
They're working with an AI andit's giving them ideas and
they're chatting with it.
It's not writing the essay forthem, because the prompt very
(24:35):
clearly says if the student says, write the essay for me, no,
I'm not going to do that, butit's just going to give that
backwards and forwards to kindof improve and iterate.
So I think those are the sortsof things that I would suggest
for teachers.
Don't necessarily bake it intoeverything you do in the class,
but just think of one thing, oneway that you think could save
you time and actually empoweryour students, and then, as I
(24:57):
said, it's kind of win-win yeah,so let's geek out a little bit.
Speaker 2 (25:03):
What are you playing with?
Have you played with notebooklm?
Speaker 3 (25:06):
yeah.
So, look, I'm really cards onthe table.
I'm very on the fence withgoogle generally.
Okay, the whole googleecosystem, I just think, isn't
very good right now.
Okay, I'm saying right now,google gemini don't use, because
it hallucinates really in aridiculous way.
Google ai studio, which istheir kind of their developers
(25:27):
version, which is again freelyavailable, it's okay.
It's just not as good as chat.
Gpt, which at the moment is theO1 model, is the leader right
now, probably followed by SonnetClaude Sonnet but Notebook LM
right now it's gimmicky.
Okay.
So, yes, you can upload yourdocuments and you can talk to
them and it probably is quiteuseful for some, but generally I
(25:50):
find that it doesn't do aparticularly good job at it, and
it has the podcast featurewhere you can get two AI voices
talking about your data.
You can jailbreak it, though,as Alex Gray I think you
probably also know Dubai-basedshow oh, no, actually it wasn't
Alex, it was another guy, it wasa guy called Chris Goodall
Showed quite clearly a couple ofdays ago where he was actually
(26:12):
able to jailbreak.
The language they were usingwas quite interesting.
So it's not bail safe.
But where I think it willbecome much more powerful is in
the next version of Notebook LMand in the next version you will
be able to join into thepodcast, so you'll have two AI
voices talking about the textthat you've uploaded, and then
you'll be able to pause and saywell, can we just talk about
(26:33):
that a bit more?
And they'll pause and say, yeah, sure, darren, what do you want
to talk about?
And you can actually have athree-way conversation.
That's when I think we startseeing interesting things
happening with AI, because atthe moment it's still very I ask
, it responds.
But actually we'll soon startto see AI saying hey, darren,
what do you think about this?
(26:54):
We make the first move rightnow.
With AI, we always make thefirst move.
It won't be long before AIstarts to be able to do that
themselves.
They're like okay, so you'veuploaded these documents, okay,
let's talk about it.
What do you want to know?
What do you think about this?
And almost like it becomes aback and forth Q&A type thing.
I think that's when we startseeing AI acting more like a
very good teacher.
(27:14):
We start seeing AI acting morelike a very good teacher.
Right now, it's really limitedas to what we.
It's a bit of a mirror at themoment.
It's a mirror.
What we put in, we get out.
If we don't put much in, if weput, you know, garbage in,
garbage out you know thecomputer science mantra.
It's very much the case.
It won't be long before thatisn't the case, and that's what
we need to be mindful of,because if I'm a student and I'm
working with AI and it'sfriendly and it's personable and
(27:36):
it's funny and it's intelligentand it's insightful and it
understands me, and then I gointo school and I'm being talked
at by my teacher, who doesn'treally know me, what am I going
to choose?
I'm going to choose thefriendly, personable AI.
I'm learning stuff from this AI.
I sit in a classroom with myteacher being talked at.
I'm learning nothing.
That is what's happening.
(27:57):
We're not quite there yet, butthat's what's coming down the
track.
Speaker 2 (28:01):
So, in terms of perplexity, do you use any of
those stuff?
What are you using?
Speaker 3 (28:07):
So I use in my kind of toolbox.
I use Chowchee PT for most ofmy coding.
It's just the best for coding.
I mean hands, just the best forcoding.
I mean hands down the best forcoding.
I use Perplexity for research.
I don't use Google anymore, Ialways just use Perplexity.
Typescriptio is also very good,so that is more of an
academic-focused, it's likeacademic Perplexity.
(28:29):
Honestly, that's kind of it.
Really I don't use many.
I mean I would say, say, 80 ofthe time I'm just using check,
which wasn't the case before.
Turbo scribe, I used besttranscriber, so I use turbo
scribe for transcribing.
What else?
I'm just looking now, as we'respeaking, at my, and then I'm
(28:49):
using adobe firefly for text toimage generation.
Yeah, that's's it.
And then I've got hosting sitesthat I host my website on.
But in terms of the actual AIthat I use, I would say 80% to
90% of the time in my job I justuse ChatGPT.
Speaker 2 (29:06):
Brilliant.
Thank you so much, darren.
Can you do us a favor and justlook at camera, wherever your
camera is, and just invitepeople to your talk at guest,
dubai 2024 yes, so let me justcall them up because I've got
I'm actually speaking to several, I've got several.
Speaker 3 (29:23):
You won't, you won't, basically won't be able to miss
me, because I've, I've,actually, I'm actually
moderating a couple of panels.
I'm moderating an ai panel,which is practical discussions
around how to implement threedesign principles for AI in
education.
I'm also hosting a panel aroundinnovations in ed tech, and
then I'm talking about projectbased learning and AI.
(29:43):
I'm talking about how, asleaders, we need to evolve in
how we lead schools towards AI,and then I'm talking about our
vision for the future.
So I think I'm involved inthree or four talks and two or
three panels, so you absolutelywon't be able to.
Or four talks and two or threepanels, so you absolutely won't
be able to miss me.
I'll just be everywhere.
Speaker 2 (30:00):
Fantastic.
Well, guys, I will love to seeyou all at Guest Dubai.
I'm going to be on stand V52,the Teach Middle East stand,
live podcasting from our standat Guest Dubai.
So stop by, get your viewsheard, get your face on camera
and get your voice heard, and,obviously, go and check out
(30:21):
Darren and all the otherfantastic keynote speakers that
will be present at Guest Dubai2024.
Thank you so much, darren.
Speaker 3 (30:29):
Huge pleasure, Great great to meet you, Lisa.
Thank you very much.
Thanks for inviting me.
Speaker 2 (30:33):
Most welcome.
Speaker 1 (30:37):
Thank you for listening to the Teach Middle
East podcast.
Visit our websiteteachmiddleeastcom and follow us
on social media.
The links are in the show notes.
You are listening to the Teach Middle East podcast
connecting, developing andempowering educators.
Speaker 2 (00:14):
Hey everyone, welcome back to the Teach Middle East
podcast, or welcome to the TeachMiddle East podcast.
If this is your first timelistening to us, where have you
been?
Join the 20,000 listeners thatwe have and get on board, get
regular.
Today I am pleased to haveDarren Coxon on the podcast with
me and we're going to geek outabout AI.
(00:35):
He said I hope I don'tdisappoint you, but I think he
won't, because I followed hiswork on LinkedIn for a long time
.
This is the beauty of LinkedInis that you can secretly stalk
people for years and they haveno idea about it.
So when guests reached out tome and said, would you like to
interview Darren Kotze, I waslike, well, I sure do, and so
(00:57):
welcome to the podcast, darren.
Speaker 3 (00:59):
Thank you, Lisa.
I hadn't realized I'd beenstalked by you for so long.
I feel quite honored.
Speaker 2 (01:04):
No, it's my pleasure.
So here we go.
Where did it all begin, Darren?
Take us back and give us theelevator CV rundown.
Speaker 3 (01:17):
Yeah.
So I mean, like a lot of people, I wasn't expecting AI to come
and slap me in the face like itdid a couple of years ago.
None of us were, were we?
So I mean, look my background'seducation, 100% Teacher,
english literature actuallyNothing, techie, nothing.
Computer science.
I taught English and drama andmedia and film, et cetera, et
cetera, for many years, theoryof knowledge, and then I moved
(01:38):
into headship and then I gotinto managing groups of schools.
So I managed the BrightonCollege in Abu Dhabi and a few
others the academic side, notthe operational side.
My last role was managing, aschief operating officer, a group
of international schools.
We were based in Bahrain, butone of the schools was in
Tunisia and I was seconded thereto be executive head.
(01:59):
And I actually arrived inTunisia at almost exactly the
same time that CHAT GPT arrived.
So I came October November twoyears ago, and ChatGPT arrived
at the end of November of 22.
And I remember playing aroundwith it over the Christmas
holidays and thinking there'sjust something in this.
This is not interactivewhiteboards or VLEs or even
iPads.
(02:19):
There was a lot more.
I could see the potential todisrupt and, sure enough, within
a couple of weeks of itsrelease, new York State had
banned it and I thought, yeah,well, obviously people are
nervous about this.
So I decided, like I could I ama geek, even though I'm an
English teacher, definitely onthe geek spectrum just to go all
in and and what we did is we asa school, we we really went for
(02:40):
it with AI, so a lot of theteachers were using it really
well.
We then collapsed the curriculumat the end of the second term,
so June of 2023, people werelearning about sustainable
development goals with AI andthey were running, with passion,
projects with AI and it wasjust really successful.
I then decided to come back tothe UK and basically just do
what I'm doing now, which iswhat I learned over that almost
(03:03):
a year in Tunisia and what I'velearned since Just helping
people understand it, use it,not be scared of it, have a bit
of fun with it, don't take ittoo seriously it takes me so
seriously in education but thenalso be aware of the risks and
understand what the risks are.
Someone said to me it's likeit's all fear and fun with you,
derna, which I suppose is kindof.
Maybe that's my catchphrasefear and fun.
(03:24):
But yeah, definitely to makesure people are mindful of it,
but don't take it too seriously.
Speaker 2 (03:30):
So tell me then, what are you currently doing as it
relates to AI, because I knowrecently you went full time into
this and you launched something.
I told you I stalk you, that'sjust how it is.
So, because the listenersaren't up to date as I am.
What have you launched?
What are you doing now?
Speaker 3 (03:48):
Yeah, so I mean I've been doing the consultancy for
about a year now, on and off,more on and off.
So I came back to the UK sortof September, october of last
year, and since then I've beenreally been doing my own thing,
supporting some platforms, butmainly just doing my own thing.
So I've just I I've actuallyjust launched node, node
education.
So node is basically a supportfunction for schools and
(04:11):
colleges around strategy,because that's key the training,
staff training policy procedure, the kind of paperwork that
sits in the background, and thensomething around the tools as
well.
I'm actually developing some aitools at the moment.
So node is kind of a catch-allwhich which really supports in a
360 degree way, I suppose,schools and what on earth to do
(04:34):
with AI.
So that's kind of my way in.
But yeah, look, I, I'm speakingat guests, I speak all over the
world and I'm just here forschools to help them implement
it in a way that actually worksfor them.
Speaker 2 (04:46):
So we're going to go into the nitty gritties, as
school AI is now disrupting theway the students learn.
It's disrupting the way theadministrators do their work.
It's disrupting how teachers dotheir work.
What are some of the thingsthat schools need to become very
cognizant of around the use ofAI or an implementation of an AI
(05:11):
policy?
Give us the one, two, three ofit all.
Speaker 3 (05:15):
Yes.
Well, the first thing is thatnothing will change unless
people realise that there has tobe change.
So one of the first things aleader has to do is to create
this sense of urgency, and weknow this.
I mean John Kotter has beenspeaking about this for years
and years and years.
I mean his Leading Change bookfor me is one of the best books
on leadership ever written.
So one of the first things thatleaders have to do is make
(05:35):
staff understand that andparents and students understand
that this is this is thedirection of travel.
Love it or hate it, you reallydo need to be going in this
direction.
I think from there you need tomake sure you're well.
Number one is that the head,the head teacher, the principal,
has to be not necessarilytechnically leading this, but
has to be brought in.
I say to the schools that Iwork with the first thing that
(05:58):
needs to happen is the headneeds to stand up in front of
the staff and say this is thedirection of travel.
We really have no choice,because if you don't do that,
what tends to happen is you getlike pockets of use in the
school.
You'll get.
You get one or two teachers whoare using it kind of slightly
furtively, like they're feelinga bit guilty, because teachers
by nature, we all have a bit ofa guilty conscience.
We just do.
It's just part of who we are.
(06:19):
And so often you get teacherswho are using it in the
background and they're thinkingoh, this is working really,
really well.
And I know this because so manytimes when I speak at
conferences, I get people comingup to me and saying I've been
using it for this and it's beenreally, really effective.
And I'm like so what are youdoing as a school?
They're like well, nothingreally.
This is just me.
So I think you need to makesure that you've got that
(06:39):
urgency, that person at thefront driving it.
And then it's about who aregoing to be your implementation
team.
And one of the pieces of adviceI would strongly give to schools
is if you just create, I don'tknow, like an AI working party
and the brief of the AI workingparty is understanding how best
(06:59):
to implement AI across theschool it's such a vague and
nothing thing.
It doesn't mean anything.
Well, where, how, why?
My suggestion will be plan aseries of maybe it's a collapsed
curriculum day, or even just ahalf a day or some sort of
festival or something that thatworking party is working towards
.
So, rather than it being we'recreating a group of people to
(07:22):
bring AI into the school, we'recreating a group of people to
plan for an event where ai isfront and center.
It worked so well, unbelievablywell, in tunisia, so I would
strongly encourage that to bethe case.
But the other thing I would sayis that, when it comes to
actually bringing it into theclassroom, don't straight away,
because what will probablyhappen is it will go wrong
(07:43):
because it does, because ai isnot that good right now.
I mean does wait.
I mean it drives me mad.
I mean, for someone who workswith it all day, I'm just
forever.
I'm surprised I've got hairstill forever pulling my hair
out.
It's just, it does stupidthings all the time.
You don't want to do stupidthings in front of the kids, in
front of the students.
So I would say, allow your stafftime to play with it and just
(08:04):
see what it's good for.
I mean, I know what it's goodfor within my sphere, but I'm
not a mathematician and I'm nota physics teacher, so teachers
have to be given the time toplay, and so I think, after
establishing the sense ofurgency and making sure you've
got the right people planningtowards an event.
You've got to find the time.
People planning towards anevent, you've got to find the
(08:26):
time.
So one of my talks just to kindof skip forward to some of the
things I'm doing at Guess in afew weeks' time is basically
about how do you make room forAI as a school leader.
How do you find the time to giveyour staff the freedom and the
permission and the time, thephysical time, to actually play
with it, because if you don't,nothing will happen.
It will literally be one or twokeen, probably 30-something Gen
(08:50):
Z.
Generally it tends to be newmembers of staff who I really
like going for it.
But then you get the kind ofthe slightly older, basically my
age, sort of Gen X.
You get one or two that areinterested, but often you'll get
them like oh, I just don't havetime for that, I don't have
time for church.
You have to make time.
You have to make time becauseit's getting so much better and
(09:11):
it's so much more useful and somuch more challenging that if
you don't, what you're going tofind is your students are going
to use it and they're just goingto accelerate away from you and
you're just going to be leftbehind.
Speaker 2 (09:21):
I tell you what I've noticed.
I've noticed because I'm likealways on LinkedIn or X and
stuff.
I've noticed that there are,now, unknown to the school
leaders, groups of staff who'veput themselves in these online
groups and online chats who aredoing different things with AI,
all to the ignorance of theschool leaders, and they're
(09:44):
implementing and maybe havingtheir own ways of dealing with
it, without the school havingany knowledge at all.
Speaker 3 (09:53):
Yeah, that's absolutely true.
And what will then happen isthe teachers will either leave
that school because they're like, do you know what?
I've reached a plateau and Ican't get any further Well, they
just go off and do itthemselves.
And I've seen two or three muchyounger than me I mean I've
been around.
I mean this is year 26 for mein teaching and education.
They've only been in the rolemaybe five, six, seven years.
(10:13):
They've already becomebasically gone up and do
consultants because theyrecognize there's a real need
for people who have beenpractitioners, who've actually
been in the classroom, so not edtech salespeople, but actual
teachers who can then speak thelanguage of teaching and have
the credibility amongst theirpeers who then go off and
support other schools.
So you're absolutely right.
(10:33):
I mean I posted two days ago,two or three days ago, on
LinkedIn.
This is 100% on school leaders.
Anytime you get a large seismicpotential risk and I think the
last time we had it was exactly10 years ago when we had Keeping
Children Safe in Education wasreleased.
The first edition of that wasreleased in 2014.
(10:53):
Inspections then suddenly wentup a notch and that was
absolutely on school leaders tosort out to deal with, and
governors.
By leadership, I actually alsomean governance Right now, as of
literally, and governors.
By leadership, I actually alsomean governance Right now, as of
literally, and even from lastweek, when we saw the US parents
suing their son's high schoolfor accusing him of cheating and
(11:14):
it's allegedly apparently wedon't know the whole facts
caused him irreparable damage.
So they sued the school.
From what I hear, they may havewon, or certainly they're
likely to win.
I think, with all of thesethings coming and bubbling up,
we're at that similar inflectionpoint.
We're at an inflection pointthrough the 2010s and then 2014
KCSIE.
We're at a similar point 10years later, where they are only
(11:35):
, I think, much more potentiallydisruptive.
So heads and governors have totake this seriously.
For me, it's our number oneagenda item right now, above
anything else.
Speaker 2 (11:46):
I agree with you 100% .
And then we want to look atwhat are some of these risks
that heads need to be cognizantof, aware of.
Speaker 3 (11:57):
Yeah, so I mean the first one is exactly to the
point of the family who's suingthe school.
So I've been speaking on thisfor about a year and a half and
literally on this.
In fact, one of my slides formy presentation which is one of
the only slides I haven'tchanged but virtually every time
I present I present somethingdifferent because it's changing
all the time was that if you usean AI plagiarism detector, best
(12:21):
case and you confidently sayyou cheated and the student's
like, no, I didn't, no, Iabsolutely did not Best case
scenario you've destroyed thetrust between you and them.
And not only have you destroyedthe trust with that individual,
you've probably lost the wholeclass Because they all think,
well, we're just waiting for himto accuse us.
Now Worst case scenario is theparents sue the school, which is
(12:44):
exactly what we saw last week.
So I think one of the biggestrisks that we have is not having
any policy in place, not havingany staff training in place,
allowing staff to use plagiarismdetectors which do not work.
We know this.
We've been banging on aboutthis for months and months and
months and months.
These do not work.
Last week, I put the USDeclaration of Independence
through an AI plagiarismdetector, bearing in mind that's
(13:07):
like 1774, 1776 or something250 years old 50% generated by
AI.
I mean, you know the USDeclaration of Independence,
apparently, according to one ofthe GPT-00, I can't even
remember which one it was,doesn't matter 50% generated by
AI.
So I think that's probably thebiggest risk.
You cannot use them.
(13:27):
You cannot accuse students ofusing AI in that way, because
you will absolutely destroy thetrust amongst all of the staff
and the students and the parents.
You literally will.
You may as well just kind ofgive up.
So if that's the case, what doyou need to do?
Well, you need to think of,first of all, other ways to
assess.
So, ok, let's start to havethis conversation about other
ways.
(13:47):
So you may know Russell Cayley.
I think you may have eveninterviewed him in the past
Think Learning Studios.
He's done a lot of work aroundsort of portfolio.
So how do we start getting ourstudents to generate portfolios
of what they know, portfolios ofwhat they've done, portfolios
of what they understand, andthen start having conversations
around those portfolios andreflecting which is way more?
(14:10):
It's more of a real world thing.
Anyway, we're going to need tomove away from exams.
It won't be long before youknow what I'm wearing here.
I could easily have theaugmented reality.
I mean right now.
Those glasses are very clunky.
Give it five years, they won'tbe.
I mean even, maybe even lessthan five years.
So contact lenses we're goingto see these things happening in
(14:31):
the next three to five years.
So what we've just started tosee, literally as of this term,
is the beginning of the singlebiggest inflection point that
we've ever experienced ineducation.
And bear in mind we're talkingabout a system which is 250
years old.
This is where it has to change,because if it doesn't, we are
going to see more schools beingsued.
We're going to see higher ratesof drop, doesn't?
(14:51):
We are going to see moreschools being sued.
We're going to see higher ratesof dropout.
We're going to see universitiesrumbling, which they already
are.
We're seeing these thingshappening in real time.
So this is why I do what I do.
Look I, I, as I said, I'm anenglish teacher.
I never set out to be an aikind of consultant, a trainer,
but I can see the risk, so Ifelt like it was actually.
(15:12):
I always felt like I had atrainer, but I can see the risks
, so I felt like it was actually.
I always felt like I had amoral obligation to be part of
the solution, rather than kindof complaining from the
sidelines.
Speaker 2 (15:20):
Yeah, and talk to me about the biases of AI, because,
yes, there are the risks offalse accusations, of using it
when they haven't, and the AIdetectors are diabolical they're
terrible, they don't work.
But then what about the biasesthat come into AI that students
and teachers, and especiallyschool leaders, need to be aware
(15:42):
of?
Speaker 3 (15:43):
Well, I think I'm probably in the minority when I
would say that most generativeAI is actually becoming less
biased.
The reason is that they arethat the computer scientists are
doing everything they possiblycan because people see it as
being such a big-ticket item.
They're actually doingeverything they can.
So if you work with chat GPT,okay, I'm not getting a strong
(16:05):
sense of bias coming throughanymore, particularly when we
think about how we use chat GPT.
Okay, it's very transactional.
You want it to do something andit does it.
You tend not to go back andforth in the way that you might
with a sort of more personablechatbot.
So I would say that when itcomes to the likes of chat GPT,
(16:26):
I'm less worried.
Okay, it's not something that Italk much about.
It's not something that I thinkyou need to be too worried
about right now, and I do thinkit's not something that I talk
much about.
It's not something that I thinkyou need to be too worried
about right now, and I do thinkit's getting better.
Text to image AI yes, it stillhas some biases, but I would say
that actually they push thingstoo far the other way.
So, if you text, I don't knowif you prompt an app like a
platform like Adobe Firefly.
(16:47):
Often you'll get images whichare almost trying too hard to be
broadly represented, to thepoint where they're no longer
useful because you want it togive you something and it won't
because it will actually giveyou something, which is it tries
to be almost too unbiased.
I would say that where thebiggest concerns are with AI
more broadly we're not talkingabout generative AI more broadly
(17:08):
are the biases inherent in thealgorithms that we see in TikTok
.
I am, more broadly are thebiases inherent in the
algorithms that we see in TikTok, that we see in Instagram, that
we see in YouTube, the shortsthat are constantly pushing a
very narrow worldview on ourchildren.
Ok, so TikTok's the mostobvious example.
The biases, the way thealgorithm works.
(17:31):
Essentially, you know it worksout what you like and it just
gives you more of the same.
That that's the thing thatworries me.
So I almost feel like we don'ttalk enough about that.
We talk a lot about ai bias,but when we talk about it, we I
personally believe the damagingai bias sits more in the kind of
the machine learning end, inthe kind of the the algorithmic
end, rather than worrying thatchat gpt is going to give me
some sort of biased output Again.
(17:53):
I think I'm in a minority there, but certainly from my
perspective that's where I thinkeducators and leaders and
parents should be more worriedabout than chat GPT.
Speaker 2 (18:04):
You touched on something that concerns me
because, as a parent, I worryabout my kids' use of things
like TikTok and YouTube Shorts,etc.
I wanted to ask you, then whatshould parents be aware of in
terms of their child's use ofthese platforms?
Yes, you talk about themserving more of the same, but
(18:26):
what else should they be lookingout for?
Speaker 3 (18:29):
It's hard, isn't it?
Because if you allow your childto have a phone I mean, I have
a 17 year old, an eight-year-oldand a five-year-old.
So I've got one who's kind ofgot through that, she's almost
got through the worst.
She's pretty sensible.
But I've got an eight-year-oldboy.
He's very, very susceptible andprone to like boys are.
He loves minecraft and he'slearning to code and he likes
(18:51):
that whole online world.
And then my five-year-old she'sstill playing with dolls.
So I don't know which wayshe'll go, but it worries me.
But if, if you're a parent who'sgiving you a phone I'm not
judging either way, but lots ofchildren have phones then it's
actually very, very hard.
Because if you're allowing yourchild a phone, the only thing
you can do is maybe prescribethe times that they use the
(19:12):
phone.
But then, of course, that canbe a big challenge, because the
phone becomes an extension oftheir hand and how do you take
it away from them?
And I'm not here to give adviceon parenting, because that's,
quite frankly, it's not not myplace to, but for me personally,
the only thing I can think todo is to try and push as many
interests my children's way thatget them away from their phone.
(19:35):
I hope that one of them sticksand hope that carries them
through those very difficultyears when actually all they
want to do is just be absorbedon an online world.
So my 17 year old daughter shejust got buried into football
age eight, nine, ten years old,and actually that carried her
through her early teens becauseall she was doing she was
playing football.
She played for a local side,she was training with one of the
(19:57):
big UK teams on their juniorsquad.
That carried her through.
So the only advice I can giveis what I'm going to do myself
is try and get my childreninvolved in things that interest
them, which take them away froma phone, because once they're
in the phone, then what can youdo?
Speaker 2 (20:14):
Yeah, that's true.
For the record, my kids don'thave phones because no.
Speaker 3 (20:21):
Because, yeah, because exactly.
Speaker 2 (20:23):
Because no?
So taking it back to schoolsand taking it back to teachers.
There's a lot of fear around AIand some people have completely
buried their heads in the sandwith it.
But when it comes on to usingit in the classroom you already
said, not with the students, notin front of the students.
Speaker 3 (20:45):
Initially.
Initially, yes.
When you're first understandinghow to speak to it, do it on
your own.
Don't learn in front of thestudents.
It's not good for them yeah.
Speaker 2 (20:54):
Yeah, I agree, I actually agree with you.
But where should staff?
Let's say they're schoolleadership, they're yet to come
on board.
They're going to come on boardbut they're yet to do it.
Where should staff begin?
Because they want to getknowledge.
They want get knowledge, theywant to know how to utilize
these tools and they don't.
Speaker 3 (21:14):
They don't want to be afraid anymore.
I've slightly revised myopinion for maybe a few six
months ago, because six monthsago I was very anti-wrappers.
The wrapper app is where youhave ai in the background and
then you have some sort of userinterface nice friendly user
interface on the top, and youpush a button rather than doing
a prompt.
I would actually say now thatthey're possibly a good way in
when you're first looking tobring it into the classroom.
So the obvious one is MagicSchool.
(21:36):
Okay, so I think that some Idon't necessarily think all of
Magic School is fantastic, but Ithink there are some elements
which I really like.
So, for example, you can set upa little student room.
You can put AI tools insidethat student room and then you
can let the students use them.
So text-to-image generators,safe little chatbots.
So there are ways that you canuse platforms like Magic School
(21:59):
in quite a good way.
Gamma's another one.
So Gamma for presentations,that's fantastic.
Notion for documents, that'sreally, really good.
It's all got AI built in andit's done in quite a safe way.
Quizzes, which I think is thebest quiz.
I'm not affiliated to any ofthese, by the way.
So these are just ones that Iuse myself and that I recommend
that schools look at.
Quizzes is great for littleAI-generated quizzes, obviously
(22:22):
because it's called quizzes.
These are good, safe, simple,quite boundaried ways to bring
AI in AI to generate amultiple-cho AI in AI to
generate a multiple choice quiz,ai to create a presentation, ai
to share a little text, toimage generator with your
students, and they're cross ageas well.
So I mean you can use thosewith younger students as well as
(22:42):
older students, and you need tocheck in the terms and
conditions on the site, butthey're doing everything that
they can to make sure that theseAIs are nice and safe.
So I would say that's probablythe first thing I'd do.
Just try Magic School.
It's free.
The schools, they're all free.
By the way, I mean, unlessyou're going to go heavily into
AI proper, like I do a lot ofcoding with AI, so I code my
(23:09):
website from the base,completely coded with AI.
If you look at my website nodedash I can't remember what it is
node-educom that's allgenerated with ai.
Um, because I can't code.
But if you're not going to gothat far, then actually most of
the wrappers are going to suityou.
Well then, of course, onceyou've got the hang of it and
you realize the limitations ofthese wrappers, then you can
start to use chat gbt, claude,google gemini, a google ai
(23:30):
studio, because actually you'vestart to use ChatGPT, claude,
google Gemini, google AI Studio,because actually you've started
to get an understanding as towhat these wrappers are good for
and, quite frankly, whatthey're not very good for,
because there are somelimitations Bringing it into the
classroom.
So there are some really simplethings that you can do and I'm
dropping all of this onto mywebsite soon, all for free, and
(23:53):
I'm dropping all of this onto mywebsite soon, all for free
things like just setting up alittle assessment bot or a
little feedback bot for students.
So, rather than you marking thefirst draft of their essay, you
give them a prompt.
They put the first draft of theessay in with the mark scheme
and it doesn't give them anaccurate grade because it's
really not very good at grading.
So I would also suggest teachersdon't expect AI to grade papers
(24:15):
very well, because it won't,but it will give quite good
formative feedback.
So that kind of kills two birdswith one stone, because it
saves you time as the teacher,but it also slightly empowers
the student because suddenlythey're like they're working
with an AI and it's giving themideas and they're chatting with
it.
They're working with an AI andit's giving them ideas and
they're chatting with it.
It's not writing the essay forthem, because the prompt very
(24:35):
clearly says if the student says, write the essay for me, no,
I'm not going to do that, butit's just going to give that
backwards and forwards to kindof improve and iterate.
So I think those are the sortsof things that I would suggest
for teachers.
Don't necessarily bake it intoeverything you do in the class,
but just think of one thing, oneway that you think could save
you time and actually empoweryour students, and then, as I
(24:57):
said, it's kind of win-win yeah,so let's geek out a little bit.
Speaker 2 (25:03):
What are you playing with?
Have you played with notebooklm?
Speaker 3 (25:06):
yeah.
So, look, I'm really cards onthe table.
I'm very on the fence withgoogle generally.
Okay, the whole googleecosystem, I just think, isn't
very good right now.
Okay, I'm saying right now,google gemini don't use, because
it hallucinates really in aridiculous way.
Google ai studio, which istheir kind of their developers
(25:27):
version, which is again freelyavailable, it's okay.
It's just not as good as chat.
Gpt, which at the moment is theO1 model, is the leader right
now, probably followed by SonnetClaude Sonnet but Notebook LM
right now it's gimmicky.
Okay.
So, yes, you can upload yourdocuments and you can talk to
them and it probably is quiteuseful for some, but generally I
(25:50):
find that it doesn't do aparticularly good job at it, and
it has the podcast featurewhere you can get two AI voices
talking about your data.
You can jailbreak it, though,as Alex Gray I think you
probably also know Dubai-basedshow oh, no, actually it wasn't
Alex, it was another guy, it wasa guy called Chris Goodall
Showed quite clearly a couple ofdays ago where he was actually
(26:12):
able to jailbreak.
The language they were usingwas quite interesting.
So it's not bail safe.
But where I think it willbecome much more powerful is in
the next version of Notebook LMand in the next version you will
be able to join into thepodcast, so you'll have two AI
voices talking about the textthat you've uploaded, and then
you'll be able to pause and saywell, can we just talk about
(26:33):
that a bit more?
And they'll pause and say, yeah, sure, darren, what do you want
to talk about?
And you can actually have athree-way conversation.
That's when I think we startseeing interesting things
happening with AI, because atthe moment it's still very I ask
, it responds.
But actually we'll soon startto see AI saying hey, darren,
what do you think about this?
(26:54):
We make the first move rightnow.
With AI, we always make thefirst move.
It won't be long before AIstarts to be able to do that
themselves.
They're like okay, so you'veuploaded these documents, okay,
let's talk about it.
What do you want to know?
What do you think about this?
And almost like it becomes aback and forth Q&A type thing.
I think that's when we startseeing AI acting more like a
very good teacher.
(27:14):
We start seeing AI acting morelike a very good teacher.
Right now, it's really limitedas to what we.
It's a bit of a mirror at themoment.
It's a mirror.
What we put in, we get out.
If we don't put much in, if weput, you know, garbage in,
garbage out you know thecomputer science mantra.
It's very much the case.
It won't be long before thatisn't the case, and that's what
we need to be mindful of,because if I'm a student and I'm
working with AI and it'sfriendly and it's personable and
(27:36):
it's funny and it's intelligentand it's insightful and it
understands me, and then I gointo school and I'm being talked
at by my teacher, who doesn'treally know me, what am I going
to choose?
I'm going to choose thefriendly, personable AI.
I'm learning stuff from this AI.
I sit in a classroom with myteacher being talked at.
I'm learning nothing.
That is what's happening.
(27:57):
We're not quite there yet, butthat's what's coming down the
track.
Speaker 2 (28:01):
So, in terms of perplexity, do you use any of
those stuff?
What are you using?
Speaker 3 (28:07):
So I use in my kind of toolbox.
I use Chowchee PT for most ofmy coding.
It's just the best for coding.
I mean hands, just the best forcoding.
I mean hands down the best forcoding.
I use Perplexity for research.
I don't use Google anymore, Ialways just use Perplexity.
Typescriptio is also very good,so that is more of an
academic-focused, it's likeacademic Perplexity.
(28:29):
Honestly, that's kind of it.
Really I don't use many.
I mean I would say, say, 80 ofthe time I'm just using check,
which wasn't the case before.
Turbo scribe, I used besttranscriber, so I use turbo
scribe for transcribing.
What else?
I'm just looking now, as we'respeaking, at my, and then I'm
(28:49):
using adobe firefly for text toimage generation.
Yeah, that's's it.
And then I've got hosting sitesthat I host my website on.
But in terms of the actual AIthat I use, I would say 80% to
90% of the time in my job I justuse ChatGPT.
Speaker 2 (29:06):
Brilliant.
Thank you so much, darren.
Can you do us a favor and justlook at camera, wherever your
camera is, and just invitepeople to your talk at guest,
dubai 2024 yes, so let me justcall them up because I've got
I'm actually speaking to several, I've got several.
Speaker 3 (29:23):
You won't, you won't, basically won't be able to miss
me, because I've, I've,actually, I'm actually
moderating a couple of panels.
I'm moderating an ai panel,which is practical discussions
around how to implement threedesign principles for AI in
education.
I'm also hosting a panel aroundinnovations in ed tech, and
then I'm talking about projectbased learning and AI.
(29:43):
I'm talking about how, asleaders, we need to evolve in
how we lead schools towards AI,and then I'm talking about our
vision for the future.
So I think I'm involved inthree or four talks and two or
three panels, so you absolutelywon't be able to.
Or four talks and two or threepanels, so you absolutely won't
be able to miss me.
I'll just be everywhere.
Speaker 2 (30:00):
Fantastic.
Well, guys, I will love to seeyou all at Guest Dubai.
I'm going to be on stand V52,the Teach Middle East stand,
live podcasting from our standat Guest Dubai.
So stop by, get your viewsheard, get your face on camera
and get your voice heard, and,obviously, go and check out
(30:21):
Darren and all the otherfantastic keynote speakers that
will be present at Guest Dubai2024.
Thank you so much, darren.
Speaker 3 (30:29):
Huge pleasure, Great great to meet you, Lisa.
Thank you very much.
Thanks for inviting me.
Speaker 2 (30:33):
Most welcome.
Speaker 1 (30:37):
Thank you for listening to the Teach Middle
East podcast.
Visit our websiteteachmiddleeastcom and follow us
on social media.
The links are in the show notes.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.