Jan 29, 2020
This week we're joined by Kate Carruthers, Chief Data and Insights Officer at UNSW, a university in Sydney with nearly 80,000 students and researchers. Kate talks about the interesting evolution of her role and her team's, from producing reports to becoming the data engineers and insights. We start by talking about data, and as we move to talking about Artificial Intelligence, Kate points out that "AI is the thing that is not yet in production", because once it's in production it's called something else (exactly as we found out when we talked to Troy Waller on the podcast, about the use of AI for accessibility, where AI wasn't the thing to focus on, it was the service it provided - like captions, dictation, text to speech etc).
Kate talks about the way to tackle AI and data problems - start with the problems of the organisation, not with the technology, and as she points out "If you start with the technology in mind, then you end up shaping the problem to fit the technology".
Kate also talks about UNSW's clear model for ownership of the data in the university - this is an important discussion, because in many cases using AI requires good organisational data, and in larger organisations it can be tricky to track down the data, and identify who can give permission for it to be used. In fact, in most AI projects, sorting out access to the data, accessing it, and tidying it up makes up 80% of more of the time and effort!. And as Kate makes clear, this isn't just about how the university uses the data, it's also about clarity on how student data cannot be used!
Finally, Kate discusses ethics and governance of data and artificial intelligence, and the work that is being done in the university to build a policy for AI use, alongside the existing clear policies on data.
In the podcast, Ray talks about the large proportion of AI projects that fail to deliver business benefit, and he conservatively talks about 60%. But there's a number of published reports and articles that put the number much higher:
________________________________________
TRANSCRIPT For this episode of The AI in Education Podcast
Series: 2
Episode: 4
This transcript was auto-generated. If you spot any important errors, do feel free to email the podcast hosts for corrections.
Hi, welcome to the AI and education podcast. I'm Dan.
And I am Ray.
Hi, Ray. How are you today?
Oh, I'm great, thanks, Dan.
Fantastic. And from the last podcast, we were going to go and look
at people who are smarter than us
and interview them on a podcast series.
Yeah. I threw a request out and it's amazing how many people
smarter than us that I managed to find.
Yeah. So, who have we got this week?
Well, this week I went to talk to Kate Kurthers from UNSW in
Sydney.
What's her role?
Uh Kate is the chief data and insights officer.
Oh, fantastic.
UNSW big university in Sydney. They are typical of universities in
Australia but maybe a bit bigger than the others. It's a very
prestigious university. They have some amazing projects going going
on. But Kate has been doing a lot of work around the data at the
university. So maybe have a listen to this Dan and then let's talk
about it afterwards.
Cool.
Today's guest is Kate. Kate from UNSW. Kate, do you want to tell us
a bit about who you are, what you do?
Sure. Well, I'm the uh chief data and insights officer here at
UNSW. and I'm also an adjunct senior lecturer in the school of
computer science and engineering.
What what does that mean you do?
So I look after all the data in the university. So I make sure that
we manage it as an asset. I establish policies and guidance in
respect of it. I also look after enterprise data warehousing,
business intelligence and analytics.
And UNSW is a pretty big university in global scales, isn't it?
Yes, it is. Over 60,000 students.
Wow. Typically you'll be at the larger end of universities around
the world. I remember one of your colleagues at a conference saying
that the school of mechanical engineering was bigger than the whole
of Stanford University which was a massive eye openener for me
because the big American universities have the big brands but
you've got the big number of students here in Australia.
Australia has really high student numbers for compared to the US.
So it's always interesting to see that. One thing we like to joke
about in the engineering faculty is that our engineering faculty is
the is larger than Imperial College London.
Wow.
Okay. So you work with data. So tell me what what kind of data are
you working with in the university?
So we conceive of data in three realms. So there's the
administrative data which is how we run the university as a
business. So all of our systems of record and things that we use to
administer students and staff. And then there's learning and
teaching which is all about learning management system. and related
technologies. And then there's research. And at the university, we
undertake research in every conceivable realm that you can imagine.
So we've got everything from HIV and clinical trial data to uh
downloads from satellites and climate change modeling and all sorts
of things.
So often we Dan and I were talking actually couple of podcasts ago
around the definition of big data and what big data means. And we
we came to this point which was about the three Vss. The volume of
data, the variety of the data and the velocity of data. So it
sounds like you've got volume of data because you're a big
university. You'll have lots of data. You've got variety of data
because if you've got stuff coming in from satellites, that's
unusual data. So you'll have all kinds of variety of data. The the
third bit around big data is the velocity. So do you think about
the data that you have as being big data? The data that I work with
in the administrative space, I don't conceive of as big data. Um,
I'll tell you a story. When I very first started here, I was
working in the engineering faculty. I was wandering around talking
to the researchers to try and work out how much data we were
dealing with just in the engineering faculty. And in the course of
one day, I stopped counting when I got to exabytes cuz I couldn't
do the math in my head anymore. And I was just like, there's a
whole lot of data here. But in the administrative realm, we're
really only dealing with gigabytes, terabytes of data, not pabytes.
One researcher in engineering was like, "Oh, can you spare up half
a pabyte? I forgot to order a half a pabyte of data storage." So,
you know, that's the realms they work in. In in the administrative
systems, it's a lot smaller. So, we use big data techniques to
analyze data, but we don't necessarily have big data.
And so, then with your team, your goal is to turn all of that data,
that bits and bites into information and and a way that it can be
used by the university.
We're really all about the insights now and my own title has
evolved that shows out how our thinking has evolved. So when I was
first appointed about six years ago, I was chief data officer and
then I became and my job hasn't changed at all in this entire time.
So I was chief data officer and then I became the chief data and
analytics officer and end of 2018 I became the chief data and
insights officer. So that's how our thinking is is is emerging and
what we've realized is my team used to just be about reporting. We
used to just generate reports for people and what we've realized is
we need to shift our role to be more of the data engineers and the
curators of well-known and well understood data sets for people to
access them.
So you're giving them the insights upon which they can then base
their decisions of what they're going to do.
Yeah.
And so BI is about bringing all of that data together and
presenting it to people in a way that's useful when it comes to
thinking about AI because everyone who's thinking about BI is
thinking about AI. It's like we've got all this data. How do we use
it? How would you apply artificial intelligence to it? I mean,
what's your perspective on AI? Because you must have a a view on
where next now you've got all your data together.
Yes, I do have a a perspective on AI and it's really interesting.
My old head of school always talks about AI as that thing which is
not yet in production. So AI is always the thing that's about to
happen and once it's in production, once it's live and people are
using it, it's always something else. So most of what most people
think of as AI in practice is really machine learning nowadays. And
there's a whole lot of interesting things happening in that space.
So what we talk about is the realm of AI, ML, bots, and related
technologies because it's all of those things. Um, and we're trying
to work out how best we can leverage those techn technologies and
also how we can govern them so that we can do it uh ethically and
prudently.
It it's interesting that point about AI is the stuff that isn't in
production yet because when we had our AI for accessibility
episode, we were looking at the ways that AI was helping in
software to provide accessibility for students. But it was
difficult to think about it as AI because it was just that thing in
the software that lets me dictate or that thing that listens and
reads out loud. what I'm looking at. You're you're absolutely
right. You don't think of that as AI. You just think of that as a
helpful thing. I don't think of Siri as being AI. I just think of
it as the thing that lets me send a text when I'm driving.
Yeah. Precisely. So, you know, I think that's a reasonable way to
approach it. And increasingly, we're going to have to think of how
best to use these technologies that are emerging.
So, do you think about a particular piece of artificial
intelligence technology and go, well, we could use it for this, or
do you start with the problems of the organization and say well how
can we help solve this problem?
I always prefer to do the latter. I always like to start with the
actual problem or the question and work your way back to the
technology because if you start with the technology in mind then
you tend to shape things to the nature of the technology and you
don't always end up meeting the organizations or the people's
actual needs.
But often technology people do start with that. Well, I've got this
thing.
I've got this hammer. I want nails. Give me nails.
My brother was a mechanic and an armorer and it was exactly right.
It's everything could be solved with a screwdriver and a hammer.
There was no other tool ever needed which worried me as an armorer
but anyway so what kind of problems are you thinking about from the
university's perspective to to say okay now I've got the data
together now I know we've got AI that can help us to do things
here's my problem statements here's the things I want to work
on
so we're really being guided by our colleagues across the
organization who are at the coldface who have problems. So we've
got colleagues who are grappling with uh teaching large classes and
needing to use technology like machine learning to understand some
of the patterns of the classes. Um we've got people who are
accessing data uh from all over the world because our students
don't just come to us in a classroom on campus anymore. They come
from everywhere. and they could be anywhere. So, making sure that
people get equitable access to things is really important. So,
there's a whole lot of things like that. Um, understanding our
students uh is really important for us and we've got longitudinal
data about our students for many years and we've never analyzed it
and we've never derived any insights in a systematic way. We've
always done it in an ad hoc way. So, that's going to be a lot of
interesting stuff for us to look had over the next little while I
suspect.
Yeah. Because I guess if you think about the data that you've got
on the students that have been on their journey with UNSW, not just
in UNSW, but afterwards, if you link together the information
you've got on their learning journey and then you can look at their
career trajectory because everyone's good enough to plot their
careers for us on LinkedIn these days, you can actually see
something that is about life impact in the university and that data
presumably can then be used to help you support current and future
students on their journeys.
And you know, things like we can start to look more deeply at
things like the assessments that we're doing and are they actually
having telling us the right things? Are they giving us the right
insights into the students or do we need to shift our assessment
models and things? It's going to be really interesting working with
our te colleagues in uh in learning and teaching space next year or
this year to look at that sort of thing.
It's going to be really interesting as Well, because I think where
in the generation that I grew up in, data wasn't as pervasive. You
know, if I if I went to a job interview, I would go to a job
interview to be assessed as a person. Whereas these days, the data
that exists about me beforehand is as evaluated before I get
anywhere near a person. And so using that data to help students on
their journey presumably is about all of the journey that they take
through the university being able to, as you say, provide equity
for students so that you can support the students that need help
most.
And I think I think that's going to be a big push I think from all
universities because we're all genuinely trying to create the best
experience for every student that we can and data can give us a lot
more insight into the finer grain of that and I think that is
really going to be big space for us over the next couple of
years.
Yeah, it's interesting to think about my background as a as a
marketeteer in the past where then we would talk about mass
personalization. It almost seems like that's the similar challenge
from a university perspective. If you've got 60,000 students, you
want to deliver a personalized learning journey for them at the
same time as managing an institution that delivers for 60,000
students.
Yeah. And and that's part of the challenge that we're facing
especially uh in in the engineering faculty. You know, we we've got
very large classes now with like 1500 students. in them. So, how do
you deliver a quality personalized learning experience to every one
of those 1500 people who are from different walks of life and you
know it's going to be interesting how we can adapt the technology
and the data to deliver that.
Well, it's interesting we've also talked about the data that other
organizations collect and how they collect it. Like whenever you
watch something on Netflix, you give it the thumbs up or the thumbs
down that allows it to put personalize the recommendations it gives
you and often we don't collect that data at the moment in
education. So I guess there's going to be more data not less data
in the future for you to grapple with.
I think increasingly uh we need to look at how how long we store
data too you know how long should we store data all of that stuff
because it's going to be expensive to store all of the data that
we're collecting and conceivably for an entire student life life
journey. So not just the four years they come to study with us but
for their entire lifetime be really interesting to think about how
we're going to do that.
The other area that I know we've talked about in the past is
governance of data, governance of AI. We're certainly thinking
strongly about ethics and we've had a bunch of conversations around
that because of the implications of using data in education and
using AI in education. There's a kind of gray line that we've got
to be careful a lot to cross. How are you handling that from a
university's perspective?
So, we have a process for internal data sharing agreements between
different parts of the organization so that the data owner gets to
explicitly approve the the use that the other party wants to put
that data to. And it's quite interesting when a few times recently
they've said no and there's some squeals of uh annoyance, but it
means that we're actually thinking through what people want to do
with the data and is that a legitimate purpose for it and it's
quite an important process to go through before you start using it.
In the past, we just was sure you can use this data whatever you
want. It was technically possible we'll just let you do it. And now
we've got this process where actually you need to think about
it.
That's that's quite a big step just even getting to the starting
block of having somebody that is the data owner because oft and
that can be quite confused in an organization. How did he get
there?
So there were some really fairly obvious people that own different
bits of the data. So the deputy vice chancellor academic kind of
owns all the students for coursework and the deputy vice chancellor
of research owns all the research students. So there were pretty
clear obvious data owners. So we started with those really obvious
ones and then started moving out into sort of increased concentric
circles
but I guess some of it might be a bit grayer. So if I think about
work around for example a scenario like forecasting student
attrition you'll want to know in some circumstances how how much
are they using the learning management system and so that would sit
in that academic sphere in
Right. Okay. And then you'd have stuff coming from the core student
information system that would sit in that space as well.
Yep.
And then you probably got finance data coming from some other place
or does that sit within
because it's related to a student does that sit within the academic
sphere as well?
It sits within the academic sphere. So, so, so there's pretty clear
ownership. It's not like in a bank where you go who owns the
customer and all the different product lines are like I own the
customer. It's
the deputy vice chancellor academic owns all the coursework
students and the deputy vice chancellor research owns all the
research students. It's pretty pretty clear. And so that that's one
of the benefits of being in university. There's those pretty clear
lines of authority. In the past though, nobody had ever asked those
people to consider the use of data. So they've delegated it down to
to people who are closer to the coldface, but we are asking them
and sometimes they don't like the answers that they get, but I
think it's a really valuable process
and I think it's and the reason I asked you digging a little bit
deeper into that is because I think that many institutions aren't
as advanced as I I was reading a story only a couple of weeks ago
around the transport data for our Opal cards in Sydney. So that's
the pre-loaded card that the equivalent of the Oyster in London and
the others. And apparently the data for that belongs to the
contractor that provides the card, not to the government. So when
the government want to do efficient city planning, they've got to
ask to be able to use the data. It doesn't belong to them. So I
think we'll probably be in similar situations in education as we're
using third party systems and the data system. in other places that
we've got to go and get access to our own data to use it.
Well, no. So, we take a lot of care when we're doing contracts to
ensure that we retain ownership of the data and that we we would
prefer to license third parties to only use deidentified student
data for instance. So, we manage that contractually and a lot of us
spend a lot of time reviewing contracts to make sure that
happens.
In my experience, I think you're on the leading edge of of doing
that. So that that's great news because it means you can then start
to use it in more constructive ways within the institution without
having all of the battles externally.
It's it's really fascinating though because you have some
conversations with these vendors and they're like no everyone else
has signed this contract. Why won't you? And we're like because we
need to protect our students rights. They give us the data for
certain purposes. We've only got the consents for certain purposes.
We don't have their consent to give you their data for you to go
and do whatever you want with it. We will give you limited and
constrained consent, not cut blanch. And and it's really
fascinating having these conversations with these vendors and
they're like, "Everybody else signed it." So that starts to move
from the governance thing into the ethics thing as well because
you're then making judgments about how the data can and can't be
used. And certainly when we sit down and talk about AI, there's a
whole heap of ethics conversations to have because there are things
that the technology will allow you to do, but you still need to ask
a question of should we do it? You know, facial recognition is one
example, but I guess there must be other scenarios where you could
do things with data that you want to be able to keep an eye on and
manage. How do you manage that from an ethical point of view?
So, we're actually working on an ethical approach to to governing
AI, ML, bots, and related technologies at the moment, and we've
assembled a bunch of researchers who were interested in the ethics
space. Um, to help us and we're we're probably going to have a
draft policy out in the next couple of months to start to frame up
that and to have a framework in which we can think about this. In
the meantime, we're just using our data sharing agreement process
to think about it on an individualized basis. So, we did a recent
proof of concept with Microsoft and LinkedIn and we ended up with
the data owners saying, "Look, I'm really not comfortable that
you're protecting the student data. adequately. So, we're going to
put it on hold until you can come up with a solution. So, that was
a real life example where the data owner just didn't feel that the
technology had been constructed in a way that protected the
students data rights.
And that's fascinating because you've got that situation where you
can very confidently and clearly talk about the data owner which I
think in many institutions wouldn't be there.
Now, I know when you created the data governance guidelines for use
within the university, you did a whole load of intern promotion to
help people within the university to become aware about how they
manage data, how they're respecting the data of other people within
the university and I know a lot of other universities have been
able to get access to that and use that to help them with their
work. Do you think the same will apply as you're going forward with
the ethics for AI as well?
I do because uh literally we've been doing a worldwide survey of
what's happening in the data ethics framework space and everybody's
got frameworks and no one's got any policies, procedures and
guidelines. So we are probably going to be the fir one of the first
to develop an actual policy about how you approach that. A lot of
really good work is happening though thinking around the
frameworks, the ethical frameworks, very little being prescriptive
and saying this is how we're going to do it in our place and this
is how we will approach it.
Well, that's going to be incredibly useful for other people to be
able to use as a a bit of a model to short circuit, but also to
understand some of the issues that you will have discussed and
debated that might have different answers in different
institutions. But knowing that the topics are there is is
useful
and and one of the important fundamentals for us for our whole data
and information governance approach is that it doesn't need to look
one way for the entire organization. It can look different in
different parts of the organization. It's the thinking about it and
going through the process that's important, not doing it in a
cookie cutter way that doesn't make sense. So that's a really
important way to think about it.
Yes. And I think that that's the same scenario that that we face.
So for example, with facial recognition, we've got a guideline that
says, well, if the facial recognition is going to use by law
enforcement, then we need to review it. But that doesn't mean the
answer each time is the same for each institution. There's a
scenario in the West Coast America where we wouldn't allow the
police force to use facial recognition because the scenario they
were going to be using it on would result in more women and more
people of color being probably tagged as false positives. But that
doesn't mean that we wouldn't let anybody continue to use facial
recognition in some scenarios. So I think it gets down to the what
is the consequence of the decision that you might make. Is it
something you could be confident with that won't have a a
disadvantage for a particular student cohort or a particular group
of students or staff or whatever it might be?
Yeah. And And facial recognition is a really good one where we
we've had some people say, "Oh, let's use facial recognition to
work out who's in class on campus in class." And pedagogically
we're kind of moving away from that as you need to be present in
class as a measure of your engagement. So there are people who
attend class by listening to lecture recordings. There are people
who never come to campus. So that starts to make that whole
discussion need to be broader about what is the purpose you're
trying to achieve and does it still make sense to us as an
organization?
Yeah, it might be you need that information as a signal to know
that this student isn't on campus as often and needs engaging in
other ways rather than the traditional attendance model. I think
we've had Dan and I have talked about this in schools where you
could replace taking the register in the morning with a student
standing in front of a camera and having their face recognized. But
my goodness, the impact on students from that could be that the
point in the day when the f where a human first talks to them and
mentions their name disappears and that would be a very bad thing
in school scenarios and I'm guess they're the same in in university
there will be scenarios where you could use data and AI to improve
the student journey but make it a less humanistic and personal
journey for them.
Yeah. And and realistically we want we we're still humans and we
want contact and we want connection and Part of our challenge is
being able to deliver that through the online channels as
effectively as we do in the classroom face to face. And that that's
kind of the challenge of the 21st century educator is is how do you
actually deliver that same kind of experience, that same kind of
immediacy and engagement you get from being face to face with
somebody through online and digital channels?
Gosh, that's a big question that I definitely don't feel we're
ready to tackle yet, but Perhaps I mean we had a great conversation
with David Kellerman a few weeks ago talking about what he was
doing around the humanistic element of personalizing learning. So
maybe that's something we should check in in in six months with to
to work out how the frameworks you're developing help you to
address those questions about how do you keep the personal touch
and the engagement for the students at the same time is using your
data and using AI more effectively.
Yep.
Brilliant. Thanks very much Kate. Okay, Dan, that was an
interesting conversation with Kate. I mean, we covered a lot of
ground in that conversation, but you know, what were the what were
the insights for you? What did you think about it?
Yeah, she had a really good focus on governance and policy and
frameworks. I thought it was really interesting the way that she
started to develop those and is thinking further on down the line
with um the actual policies rather than just the frameworks and
general guidelines in a large institution like a university. I
think that's very uh tricky to drive consistency and you can get
gaps in your security posture. The way this they tackling this
problem in UNSW seems really interesting and there was a big focus
on that governance element. So, and even bringing in that global
practice and sharing what they've done was really interesting as
well. So, I think that really stood out for me. There was a lot on
governance.
Yeah, there was a strong message there, wasn't it? which was I've
looked at other universities tried to understand what other people
are doing and and many people are being reactive in that space and
haven't yet developed the policies. So I think it's really shows a
maturity of thinking there around we're going to start building
those policies as we start to develop what we're doing not
afterwards which is often what happens you know policies as you
know policies tend to be developed as a knee-jerk to something
going wrong
and so you know getting ahead of that game they've done great work
around that in the data sphere, but now thinking about it in the AI
sphere, which is potentially a lot more complex, getting that work
done early, bringing the consensus on on board of people around the
university, clarifying the ownership of the data, clarifying what
can and can't be done, you know, that conversation around who owns
the data, not just the data that the university has, but also the
data that's in the systems that the university is using. That's
really, you know, that's been was a really good part of the
conversation that I think I don't think I would have been able to
have with many people that we talked to around the use of AI.
Yeah. And and the other thing that came out connected to that was
the value of the projects that they're doing because obviously
there's been some research that we were talking about in previous
episodes around the value of what has actually been done with the
AI projects and whether that actually you know some of the projects
what was the research something like 60%
yeah something around 60% of the projects don't deliver value back
to the organization in a way that they expected it to.
Yeah. And she's she's starting with that question rather than the
actual data which is really interesting. you know, high level
answering some of those really thorny questions in their university
context.
Yeah, it was a bit embarrassing really because Kate's going, "Yeah,
but you start with the business problem, then you think about the
technology." And of course, you know, we think about the technology
first. We've got we've got a podcast that is all about artificial
intelligence. You know, here's AI. What problem can we solve? But
absolutely right, the the point that Kate made really clearly and I
think it comes across in the way we think and the conversations
we're having is what is the problem we're trying to solve and stay
fixated on solving the problem rather than going well the techn
ology can do this so let's do this and that I think comes back to
the root of if AI projects fail that's no different from other
technology projects if you look at the reports a lot of government
technology projects fail because they get focused on the technology
and you kind of forget that what you're really trying to solve is a
is a complex problem a complex problem of student graduation or
students getting through year 12 or students doing as well as they
possibly can in their learning and so you have to stay fixated on
that solving the problem rather than getting sucked into the joy of
data science for example.
And and the other thing that that jumped out to me as well was one
of the examples towards the end there when you mentioned facial
recognition um brought that up because it it did resonate because
I've seen a lot of uh schools and unise think about facial
recognition for attendance purposes and then the way she threw that
into well actually is that what we want for online courses? Are we
actually using technology for solutions that we might need? today
but maybe won't be relevant for tomorrow. So that was really
interesting pedagogically as well when you're thinking about
this.
Yeah. And also from a a data ethics and governance point of view if
you start with a bit of work on a project that is about oh well I'm
going to do facial recognition to take a register to know who's
attended that's only one way of looking at it. Then there's also
the well which students are online which students are all kinds of
different aspects to that project as you say say to think in the
world say of a university where it's no longer about bums on seats
in a lecture hall. and a proportion of students are remote and a
proportion of students are never going to log onto a system and a
proportion of students are just going to look at things in a book
and how do you recognize all of those things and that's where that
Kate didn't use the word but the way I see it as building a broad
coalition across the institution of people that are interested in
the project and have a view to put into that. So for example the
governance piece the people who own the data the people are able to
do the analytics on the data the people that maybe own the systems
that the data is coming from building that ition so that everyone's
bought into it. So that you're not just solving today's problem
with today's data, but you're thinking about how you solve that
problem as the organization or the learning methodology or what you
are trying to achieve starts to change. You're therefore building a
coalition of people in the institution that are interested in
solving the problem and are invested in that rather than two or
three people going off and doing a bright spark project that then
doesn't become gives a good foundation. doesn't it for for all
projects to come. So once you get that foundation sorted then
everything else can the other the other element that she mentioned
as well was about the from the students point of view equity of you
know students who may be global or even lecturers or academics who
may be globally kind of connected in so they've got to have an
equitable use of AI and I suppose that would cover inclusivity as
well as just geographical location and then the other thing that
jumped out to me as well when she talked about storage because
there is an abundance of data and we've explored on lots of other
podcasts about how much data can be stored, what data is available,
what data can be stored. You highlighted in the last one about
there'll be a more discerning use of data and people thinking about
what data they'd use more. And then she brought up in in this
particular podcast about the fact that from a IT point of view, she
also thinks about the storage costs and where that might be. And
then also there's that security knock on effect of if you've got 20
30 years worth of data of students, lifelong learning and all that,
then what happens if you get data breaches and things?
Yeah, maybe the There's an argument that we need to think more
carefully about live data and data that is archived and put away
somewhere else where it isn't available to anybody to be able to
use, you know, and that's this isn't just a AI thing. This isn't a
data thing. This is a security thing. If you've got 20 years of
data in a live system, that 20 years of data can be accessed or
used in a harmful way. So, maybe it is one of the things we're
going to have to think about because we need the Marie Condo
approach to data. Does this data bring me joy?
Yeah, but it On the other hand, you know, like I've been at the
brunt of that. You my teacher training degree in the university in
Swansea, they had a fire and they lost some of the transcripts. So,
you know, it is they do need, you know, actually that that does
have a hindrance on my application for teaching in Australia
because I can't find my actual written documentation.
You they've lost the piece of paper that says you were at the
university.
Well, they they know I was there, but they don't know what I
studied. And Australia need to know what exact units you studied.
So it it becomes um and we have the same things with GPS needing to
know exactly what modules they've studied in the university.
I find that incredible. I always think I'm lucky to have not been
growing up in the Facebook age. So nobody knows about my teenage
life because it wasn't constantly recorded. But the downside of
that is the institutions don't have that data digitally either.
Yeah. No, absolutely. So So that was an excellent podcast. I'm
really looking forward to the next one.
Well, you know, we said we could go out and find people smarter
than us. That that that was great because, you know, I felt that we
were learning listening to that conversation and so you know I'm
sure we're going to go out and find some even smarter people than
us there.
Yeah. So that's great and it's uh there's some real value in doing
those conversations.
Thanks.
Thanks Ben and bye.