Generative AI in EAP - where are we now?

WEBVTT
00:00:06.919 --> 00:00:11.480
Hi and welcome to All Things EAP, a podcast produced
00:00:11.480 --> 00:00:13.839
by practitioners across the globe and hosted
00:00:13.839 --> 00:00:17.140
by Barleap. It provides a fun and friendly look
00:00:17.140 --> 00:00:20.059
at every aspect of EAP, helping you keep up with
00:00:20.059 --> 00:00:22.480
the latest developments and building an inclusive
00:00:22.480 --> 00:00:25.949
and equitable community around the world. Following
00:00:25.949 --> 00:00:28.649
the Barloop values, we hope to be collaborative,
00:00:29.010 --> 00:00:32.530
professional, developmental and inclusive. Each
00:00:32.530 --> 00:00:35.630
episode will dive into an aspect of EAP with
00:00:35.630 --> 00:00:38.869
discussions, reflections and insights from practitioners
00:00:38.869 --> 00:00:43.170
in different contexts. Whoever and wherever you
00:00:43.170 --> 00:00:46.090
are, we hope you find this podcast informative
00:00:46.090 --> 00:00:56.130
and insightful. Hello and welcome to another
00:00:56.130 --> 00:00:59.109
episode of All Things EAP. I'm Laura Richards
00:00:59.109 --> 00:01:02.750
from the University of Leeds in the UK. And I'm
00:01:02.750 --> 00:01:05.469
Xiaoqing Bi from Xi 'an Jiao Tong Liverpool University.
00:01:06.109 --> 00:01:09.849
This episode is facilitated by me and Paul Breen
00:01:09.849 --> 00:01:14.049
from UCL. Today we're diving into a question
00:01:14.049 --> 00:01:16.909
that's at the centre of so many conversations
00:01:16.909 --> 00:01:20.790
in EAP and higher education right now. Is generative
00:01:20.790 --> 00:01:24.299
AI inevitable? And more importantly, should we
00:01:24.299 --> 00:01:27.120
accept that narrative? It's a brilliant question,
00:01:27.239 --> 00:01:30.060
Laura, and it's not just about whether we use
00:01:30.060 --> 00:01:32.920
AI or not. The panel really challenged the idea
00:01:32.920 --> 00:01:36.359
that we're somehow passive in all this. The conversation
00:01:36.359 --> 00:01:40.519
went much deeper into politics, values, and who
00:01:40.519 --> 00:01:43.480
actually gets to set agenda in our classroom
00:01:43.480 --> 00:01:46.340
and institutions. Having listened to the discussion,
00:01:46.480 --> 00:01:48.540
I'm really keen to hear what struck you about
00:01:48.540 --> 00:01:51.579
it. But first, let me introduce our panel. We
00:01:51.579 --> 00:01:54.840
have Ayanna Private Goldstein from the UCL Academic
00:01:54.840 --> 00:01:57.939
Communication Center, Mike Groves from Lingnan
00:01:57.939 --> 00:02:00.620
University in Hong Kong, and he is a returning
00:02:00.620 --> 00:02:03.469
guest on the podcast. We also have Klaus Munt
00:02:03.469 --> 00:02:05.989
from the University of Nottingham in the UK and
00:02:05.989 --> 00:02:08.849
Julia Montanari from the Open University also
00:02:08.849 --> 00:02:12.789
in the UK. Four really distinct voices actually,
00:02:12.909 --> 00:02:16.449
they come at this from different angles. Pedagogy,
00:02:16.750 --> 00:02:20.069
policy, the ethics of technology, the history
00:02:20.069 --> 00:02:23.330
of education, but they all push back against
00:02:23.330 --> 00:02:27.650
the idea that we have no choice. Yeah, so what
00:02:27.650 --> 00:02:30.050
kind of things came up for you as you were listening?
00:02:32.670 --> 00:02:35.189
Well, there's a really interesting discussion
00:02:35.189 --> 00:02:38.090
about friction, you know, the difficulty and
00:02:38.090 --> 00:02:41.789
effort involved in learning and how AI can remove
00:02:41.789 --> 00:02:44.969
that friction with a couple of clicks. But if
00:02:44.969 --> 00:02:47.770
it removes friction, do we remove the learning?
00:02:48.409 --> 00:02:53.169
And what about the question of truth? Oh, that
00:02:53.169 --> 00:02:56.430
was powerful. And Julia made the point that the
00:02:56.430 --> 00:02:59.789
system aren't really designed to tell the truth.
00:03:00.009 --> 00:03:03.530
Their job is to generate text that sounds plausible.
00:03:04.050 --> 00:03:06.110
That's a really fundamental difference from the
00:03:06.110 --> 00:03:08.710
kinds of technologies we've adapted to in the
00:03:08.710 --> 00:03:12.110
past. And yet we are being asked to integrate
00:03:12.110 --> 00:03:14.849
them into our teaching, our assessment and our
00:03:14.849 --> 00:03:17.770
learning outcomes, without really stop to think
00:03:17.770 --> 00:03:21.810
about what that actually means. So it's not just
00:03:21.810 --> 00:03:25.310
how do we use this tool? It's asking what we
00:03:25.310 --> 00:03:29.370
actually value in education? Yeah, exactly. They
00:03:29.370 --> 00:03:32.409
kept coming back to that. If we center learning,
00:03:32.610 --> 00:03:35.590
if we center values like honesty and integrity
00:03:35.590 --> 00:03:38.669
and student agency, the conversations about AI
00:03:38.669 --> 00:03:41.310
looks completely different. You know, this feels
00:03:41.310 --> 00:03:43.909
particularly timely because we've just hosted
00:03:43.909 --> 00:03:46.419
a Bali professional issues meeting. here at the
00:03:46.419 --> 00:03:49.819
University of Leeds. And the topic was EAP and
00:03:49.819 --> 00:03:53.479
the Academy in the Age of Gen. AI. So this episode
00:03:53.479 --> 00:03:56.000
is really part of that ongoing conversation we're
00:03:56.000 --> 00:03:59.400
having as a sector about how to respond thoughtfully
00:03:59.400 --> 00:04:01.879
to these technologies. Absolutely. And I think
00:04:01.879 --> 00:04:04.740
the panel gives us some really useful frameworks
00:04:04.740 --> 00:04:07.340
for thinking about it. They're not just saying
00:04:07.340 --> 00:04:10.740
don't use AI. They're saying ask harder questions
00:04:10.740 --> 00:04:14.689
first. Exactly. And if you'd like to catch up
00:04:14.689 --> 00:04:17.410
on that PIM and other Barleap video resources,
00:04:17.769 --> 00:04:20.589
please visit the Barleap YouTube channel. Well,
00:04:20.689 --> 00:04:22.529
I think our listeners are going to find this
00:04:22.529 --> 00:04:27.990
really valuable, so let's get into it. Hi, my
00:04:27.990 --> 00:04:30.089
name is Claus Wundt. I'm the Deputy Director
00:04:30.089 --> 00:04:32.189
for Education and Student Experience in the School
00:04:32.189 --> 00:04:34.509
of Education at the University of Nottingham.
00:04:34.709 --> 00:04:37.750
I'm Mike Groves. I'm Senior Lecturer at the Centre
00:04:37.750 --> 00:04:40.319
for English and Additional Languages. at Lingnan
00:04:40.319 --> 00:04:43.379
University in Hong Kong. Hello, I'm Iana Privat
00:04:43.379 --> 00:04:45.860
-Golstein. I'm associate professor and head of
00:04:45.860 --> 00:04:48.300
the UCL Academic Communication Centre. And I
00:04:48.300 --> 00:04:51.300
also work within UCL on policy and guidance related
00:04:51.300 --> 00:04:54.300
to use of generative AI. And hello from me, I'm
00:04:54.300 --> 00:04:57.120
Julian Wallenari. I'm a lecturer in professional
00:04:57.120 --> 00:05:00.120
academic communication at the Open University's
00:05:00.120 --> 00:05:04.079
graduate school. And I'm academic lead for generative
00:05:04.079 --> 00:05:09.879
AI at doctoral level. Thanks, everyone. So we're
00:05:09.879 --> 00:05:12.560
just going to start quite broadly here. How far
00:05:12.560 --> 00:05:15.800
do we actually accept that generative AI is now
00:05:15.800 --> 00:05:18.480
inevitable in education? Or is that something
00:05:18.480 --> 00:05:22.100
we should even be accepting? So I'd be very happy
00:05:22.100 --> 00:05:25.660
to start with that question. I think that the
00:05:25.660 --> 00:05:28.759
more we accept the narrative that it is inevitable,
00:05:28.879 --> 00:05:31.939
the more inevitable it becomes. It's a sort of
00:05:31.939 --> 00:05:35.040
hegemonic idea and it comes to feel like common
00:05:35.040 --> 00:05:38.819
sense, like a given thing. sort of about how
00:05:38.819 --> 00:05:41.180
a cultural and ideological process connects with
00:05:41.180 --> 00:05:43.899
an economic base, with very much a private sector
00:05:43.899 --> 00:05:46.339
driven set of developments. And I do think that's
00:05:46.339 --> 00:05:50.339
what we're seeing in education. Since 2023, I've
00:05:50.339 --> 00:05:52.779
been on many cross institutional and faculty
00:05:52.779 --> 00:05:55.980
working groups seeking to frame our approach
00:05:55.980 --> 00:05:59.980
and policy. And despite the range within those
00:05:59.980 --> 00:06:03.439
groups of informed dissent, as well as critical
00:06:03.439 --> 00:06:06.889
support, Generally speaking, in the outcomes
00:06:06.889 --> 00:06:09.430
from those groups, there's always an implication
00:06:09.430 --> 00:06:13.709
that uncertainty or non -use are positions to
00:06:13.709 --> 00:06:18.170
be overcome or kept in scope rather than engaged
00:06:18.170 --> 00:06:23.029
with. And I do think that it is not about all
00:06:23.029 --> 00:06:25.529
or nothing within the education sector. We could,
00:06:25.529 --> 00:06:27.930
for example, work together more than we're currently
00:06:27.930 --> 00:06:30.850
doing to set the agenda for tools used in education
00:06:30.850 --> 00:06:34.199
to share research. to critique. Yeah. I think
00:06:34.199 --> 00:06:39.040
the point about this is a private sector technology
00:06:39.040 --> 00:06:42.459
being sold into schools and universities is a
00:06:42.459 --> 00:06:45.600
really, really important one. And EduTech is
00:06:45.600 --> 00:06:49.000
such a massive business and it doesn't always
00:06:49.000 --> 00:06:51.180
deliver what it promises in terms of educational
00:06:51.180 --> 00:06:55.060
outcome. Where I would differ is that I would
00:06:55.060 --> 00:06:58.800
think, I mean, AI is inevitable in society. and
00:06:58.800 --> 00:07:00.819
therefore it's inevitable, I think, in education.
00:07:00.959 --> 00:07:04.339
What isn't inevitable is the ubiquity of it.
00:07:04.720 --> 00:07:10.920
I would like to see pedagogy develop where we
00:07:10.920 --> 00:07:14.720
learn to turn off the AI. We learn when to turn
00:07:14.720 --> 00:07:18.160
it on and when not to use it as well. I'm just
00:07:18.160 --> 00:07:20.540
going to say... that I agree with that. The one
00:07:20.540 --> 00:07:22.759
point that I'll disagree with you, Mike, is that
00:07:22.759 --> 00:07:25.240
actually AI isn't edutech and it's never been
00:07:25.240 --> 00:07:27.759
designed to be edutech, I think, but it is used
00:07:27.759 --> 00:07:31.420
as such. Everything else chimes very well with
00:07:31.420 --> 00:07:33.459
me and also especially what you said, Ayana,
00:07:33.519 --> 00:07:36.079
the idea that as we're in education and we're
00:07:36.079 --> 00:07:37.899
constantly outward looking, our institutions
00:07:37.899 --> 00:07:39.980
always, what does the sector do? What does that
00:07:39.980 --> 00:07:42.199
want? What does that want? And so on. And I think
00:07:42.199 --> 00:07:45.120
we need to set our own agenda somehow and remain
00:07:45.120 --> 00:07:47.660
in control of that agenda as well. Because we
00:07:47.660 --> 00:07:51.100
do this for a reason, I think. So I don't think
00:07:51.100 --> 00:07:54.740
we need to or should accept the inevitability
00:07:54.740 --> 00:07:57.500
or the ubiquity of the technology. I think we
00:07:57.500 --> 00:08:01.160
should. accept that it exists, but we also should
00:08:01.160 --> 00:08:04.459
find ways of using and, as you say, Mike, very
00:08:04.459 --> 00:08:08.560
importantly, deciding not to use it for our purposes.
00:08:09.040 --> 00:08:11.939
I agree, Klaus. It is inevitable in the sense
00:08:11.939 --> 00:08:16.420
that it's there, it's real, but we can push back
00:08:16.420 --> 00:08:19.660
against that inevitability by adopting a critical
00:08:19.660 --> 00:08:22.220
stance so you can accept the reality of things.
00:08:22.910 --> 00:08:25.230
They're beyond us. They're beyond our individual
00:08:25.230 --> 00:08:27.329
agency. Sometimes they're beyond our collective
00:08:27.329 --> 00:08:32.370
agency. But we are critical beings. We're cognitively
00:08:32.370 --> 00:08:35.350
critical beings. We're reflective beings. We
00:08:35.350 --> 00:08:40.970
do have agency. Our agency is related to the
00:08:40.970 --> 00:08:43.809
structures that we're in. It's determined to
00:08:43.809 --> 00:08:47.730
some extent, but it's not anchored inevitably
00:08:47.730 --> 00:08:51.269
to the structures and the history of ideas and
00:08:51.269 --> 00:08:54.950
the history of... Knowledge and the history of
00:08:54.950 --> 00:08:58.669
social movements shows that when you do act as
00:08:58.669 --> 00:09:02.009
a collective, as Ayanna was saying, are we working
00:09:02.009 --> 00:09:06.210
together? You can achieve transformation and
00:09:06.210 --> 00:09:09.169
you can achieve change. And it's not about pushing
00:09:09.169 --> 00:09:11.870
one narrative over another, but it's ensuring
00:09:11.870 --> 00:09:15.490
that one narrative isn't over determined. There's
00:09:15.490 --> 00:09:18.549
a lot of evidence from critical scholarship in
00:09:18.549 --> 00:09:22.309
generative AI. that techno -determinism and techno
00:09:22.309 --> 00:09:25.470
-fatalism are embedded by design, right? The
00:09:25.470 --> 00:09:28.350
big tech companies don't want us to be critical
00:09:28.350 --> 00:09:32.129
and push back, so why should they create systems
00:09:32.129 --> 00:09:36.429
that allow for that push back? Which raises another
00:09:36.429 --> 00:09:38.950
big issue for me and it's the kind of the trope
00:09:38.950 --> 00:09:41.710
of the human in the loop. You know, if we prompt
00:09:41.710 --> 00:09:44.629
it in the right way or if we kind of choose to
00:09:44.629 --> 00:09:48.320
opt out, then that's all fine. But I suppose
00:09:48.320 --> 00:09:50.940
I would push back against that as well because
00:09:50.940 --> 00:09:54.100
the technology is not designed for us to push
00:09:54.100 --> 00:09:57.539
back. It's convenient for the big tech companies
00:09:57.539 --> 00:10:01.360
to feel that the responsibility is on us to make
00:10:01.360 --> 00:10:05.879
decisions. And actually, that's a bit of a myth
00:10:05.879 --> 00:10:08.759
because how are you supposed to make decisions
00:10:08.759 --> 00:10:11.720
if you don't know what you're prompting? And
00:10:11.720 --> 00:10:13.379
that's the whole point why you're prompting these
00:10:13.379 --> 00:10:14.980
things because you don't know stuff. So how can
00:10:14.980 --> 00:10:17.860
you critique? what comes out of it. You make
00:10:17.860 --> 00:10:20.340
a very interesting and very important point there
00:10:20.340 --> 00:10:23.299
and that is that technology in general I would
00:10:23.299 --> 00:10:25.960
say and this particular technology especially
00:10:25.960 --> 00:10:29.940
is designed to keep you engaged with it and disengage
00:10:29.940 --> 00:10:32.860
with other things. Hence the ubiquity of summaries
00:10:32.860 --> 00:10:34.539
which don't actually want you to read for three
00:10:34.539 --> 00:10:36.379
hours you know that kind of thing and that's
00:10:36.379 --> 00:10:38.340
that's been a phenomenon for a long time that
00:10:38.340 --> 00:10:41.490
has nothing to do with AI. But I would say It
00:10:41.490 --> 00:10:44.450
seems a bit too simple to point at the big tech
00:10:44.450 --> 00:10:47.309
companies. While of course they do this by design,
00:10:47.330 --> 00:10:50.210
I would agree to that. But if we talk about agency,
00:10:50.370 --> 00:10:53.110
that it is also on me the user or the non -user
00:10:53.110 --> 00:10:56.389
to say I want to use this for a purpose or I
00:10:56.389 --> 00:10:59.070
do not want to do that. So I wouldn't remove
00:10:59.070 --> 00:11:02.090
that agency and that responsibility from each
00:11:02.090 --> 00:11:05.610
individual because then we're basically doing
00:11:05.610 --> 00:11:07.870
the same thing somehow. We're kind of taking
00:11:07.870 --> 00:11:11.279
the responsibility away from everyone. I agree
00:11:11.279 --> 00:11:14.240
wholeheartedly with so much of what Julia said
00:11:14.240 --> 00:11:18.559
and Klaas's response also, but I do think that
00:11:18.559 --> 00:11:21.299
we have to consider how much are we empowering
00:11:21.299 --> 00:11:24.659
students to make those decisions? How much understanding
00:11:24.659 --> 00:11:28.100
do they have? It's an awful lot of responsibility
00:11:28.100 --> 00:11:30.259
to put on a student, especially if we're setting
00:11:30.259 --> 00:11:33.700
up a structure which I suppose pushes them towards
00:11:33.700 --> 00:11:37.750
potentially choosing to use these tools. making
00:11:37.750 --> 00:11:39.690
them feel that they're missing out if they're
00:11:39.690 --> 00:11:42.389
not using them, making them feel like they're
00:11:42.389 --> 00:11:45.570
disadvantaged, or setting assessments which presuppose
00:11:45.570 --> 00:11:48.470
use of these tools. I think it's our responsibility
00:11:48.470 --> 00:11:51.590
as well as educators and as an institution to
00:11:51.590 --> 00:11:53.690
move away from a sort of instrumentalist approach,
00:11:53.809 --> 00:11:55.330
you know, how do you use these tools, not that
00:11:55.330 --> 00:11:57.950
that's not important in certain contexts, to
00:11:57.950 --> 00:12:00.250
thinking about students to make those decisions
00:12:00.250 --> 00:12:02.330
that you're talking about, Klaus. I suppose,
00:12:02.490 --> 00:12:04.710
making sure that they understand how the tech
00:12:04.710 --> 00:12:06.950
works before they think about how to use it,
00:12:06.950 --> 00:12:09.029
I suppose, and what the impact of using those
00:12:09.029 --> 00:12:13.750
tools are. I think it's an interesting point.
00:12:14.090 --> 00:12:16.549
And the thing that I keep coming back to is thinking
00:12:16.549 --> 00:12:20.570
about friction, because this is a relative, I
00:12:20.570 --> 00:12:23.470
haven't heard this term in terms of education
00:12:23.470 --> 00:12:26.350
until quite recently, but this idea of effort
00:12:26.350 --> 00:12:31.289
and difficulty being key to learning. the way
00:12:31.289 --> 00:12:34.289
the AI just can take away that friction. You
00:12:34.289 --> 00:12:37.169
can create an essay that would have taken a day
00:12:37.169 --> 00:12:39.529
or two days and you can, on the surface at least,
00:12:39.750 --> 00:12:43.830
produce it with a couple of clicks. I mean, if
00:12:43.830 --> 00:12:46.870
my undergraduate years are anything to go by,
00:12:46.870 --> 00:12:48.629
I wasn't a very good undergraduate student, I'll
00:12:48.629 --> 00:12:53.470
admit it now. I didn't really see the point of
00:12:53.470 --> 00:12:55.889
a lot of the stuff that I was writing. I wasn't
00:12:55.889 --> 00:12:57.990
engaged with a lot of stuff that I was doing.
00:12:58.090 --> 00:13:02.460
So if AI had been... available I would probably
00:13:02.460 --> 00:13:05.940
have used it in a way that minimized friction
00:13:05.940 --> 00:13:09.899
for me getting through the course and therefore
00:13:09.899 --> 00:13:12.700
minimize my learning. Instead I used other ways
00:13:12.700 --> 00:13:14.419
to minimize the friction and minimize the learning
00:13:14.419 --> 00:13:18.200
that I'm not going to go into now. I completely
00:13:18.200 --> 00:13:21.539
agree with that and like that metaphor of friction.
00:13:21.860 --> 00:13:24.919
And I guess what it raises for me is the bigger
00:13:24.919 --> 00:13:27.779
question of what must the world be like or what
00:13:27.779 --> 00:13:30.899
must education be like for that friction to not
00:13:30.899 --> 00:13:34.240
be okay, for us to not sit with the trouble,
00:13:34.720 --> 00:13:38.779
as Sarah Ahmed might say. And to add to the frictions
00:13:38.779 --> 00:13:41.100
that you've mentioned, I would add other frictions
00:13:41.100 --> 00:13:43.840
such as getting rid of imperfection. getting
00:13:43.840 --> 00:13:48.320
rid of bumpy, unsmooth reads, to sort of refer
00:13:48.320 --> 00:13:50.960
to Joan Turner's work on the smooth read, the
00:13:50.960 --> 00:13:54.659
kind of ideological assumption that is actually
00:13:54.659 --> 00:13:57.779
the aim of writing, for example. So I think the
00:13:57.779 --> 00:14:01.059
bigger question is, what must it be like for
00:14:01.059 --> 00:14:04.139
that to be a problem, for all these frictions
00:14:04.139 --> 00:14:07.419
to be a problem? Class, I completely know that
00:14:07.419 --> 00:14:08.960
all these things have been happening forever.
00:14:09.019 --> 00:14:11.000
And these issues have come up again and again,
00:14:11.000 --> 00:14:15.820
way before any of the generative AI summarizing,
00:14:16.080 --> 00:14:18.700
turn it in, all the rest of it. You know, essay
00:14:18.700 --> 00:14:21.720
mail, it's all been done before. We've been there
00:14:21.720 --> 00:14:26.000
before. But I think what these social technologies
00:14:26.000 --> 00:14:32.659
do is they constantly flag up the big question
00:14:32.659 --> 00:14:36.220
of what... world do we want to live in? What
00:14:36.220 --> 00:14:39.220
are the values that we are subscribing to? And
00:14:39.220 --> 00:14:42.159
I think what generative AI is doing, and it is
00:14:42.159 --> 00:14:45.039
to do with big tech, because it was never an
00:14:45.039 --> 00:14:49.100
educational imperative to do this. It was imposed
00:14:49.100 --> 00:14:53.159
from the top down. And now we're all struggling
00:14:53.159 --> 00:14:56.519
to define education in terms of what the machines
00:14:56.519 --> 00:14:59.259
are telling us education is. So yes, it has all
00:14:59.259 --> 00:15:03.710
been there, but I think the sort of human propensity
00:15:03.710 --> 00:15:07.070
to constantly critique is there and I think we
00:15:07.070 --> 00:15:12.600
should never ever stop questioning it. If I may
00:15:12.600 --> 00:15:14.580
go really old school now for a second, I mean
00:15:14.580 --> 00:15:17.139
this idea of friction, difficulty in learning,
00:15:17.519 --> 00:15:20.399
I can trace that easily back to ancient Confucianism.
00:15:20.679 --> 00:15:23.840
The idea that learning is difficult, achievement
00:15:23.840 --> 00:15:26.600
is difficult, self -improvement if you like,
00:15:26.659 --> 00:15:28.860
and I'm not talking about perfection here. The
00:15:28.860 --> 00:15:31.279
perfection narrative is linked to the deficit
00:15:31.279 --> 00:15:33.779
model. I don't care about a typo really, but
00:15:33.779 --> 00:15:37.259
the whole idea of difficulty in learning It is
00:15:37.259 --> 00:15:39.379
a recurring theme and I think it is a recurring
00:15:39.379 --> 00:15:41.559
theme for a reason. We learn with challenges.
00:15:41.700 --> 00:15:44.360
We remove the challenge, we don't learn. So as
00:15:44.360 --> 00:15:47.159
you said, Mike, you remove the friction and there's
00:15:47.159 --> 00:15:50.480
nothing. Oh, there's very little. And the other
00:15:50.480 --> 00:15:54.440
thing that's gone back to the beginning of human
00:15:54.440 --> 00:15:58.740
civilization is technological change. And with,
00:15:58.799 --> 00:16:02.259
I can remember teaching in the nineties and we
00:16:02.259 --> 00:16:04.909
got an overhead projector. and people were talking
00:16:04.909 --> 00:16:06.789
about using the overhead projector as a learning
00:16:06.789 --> 00:16:10.450
technology. I can remember when students would
00:16:10.450 --> 00:16:13.070
copy and paste from Wikipedia and not take all
00:16:13.070 --> 00:16:16.649
the hyperlinks out. We adapted, we adopted, we
00:16:16.649 --> 00:16:19.570
changed parts of our praxis and we got better
00:16:19.570 --> 00:16:24.850
at it. I do think there's a question of why is
00:16:24.850 --> 00:16:28.169
it suddenly... Education pedagogy has evolved
00:16:28.169 --> 00:16:30.710
and developed with technology. I don't think
00:16:30.710 --> 00:16:33.139
it is. I just think because we're in the eye
00:16:33.139 --> 00:16:35.860
of the storm. Now, I'm not saying it's not important,
00:16:35.960 --> 00:16:39.899
it's not difficult, but I'm also suggesting that
00:16:39.899 --> 00:16:42.460
we've been here before. I couldn't agree more.
00:16:42.620 --> 00:16:44.980
I think perhaps what's different this time is
00:16:44.980 --> 00:16:47.320
that whereas many of the developments over the
00:16:47.320 --> 00:16:50.980
last decade have impacted primarily on students
00:16:50.980 --> 00:16:52.779
with English as an additional language in that
00:16:52.779 --> 00:16:54.940
particular community in terms of thinking about
00:16:54.940 --> 00:16:57.379
the impact of sort of machine translation, for
00:16:57.379 --> 00:17:00.889
example. or synthesis and summary, what we're
00:17:00.889 --> 00:17:03.629
seeing now is that this is impacting on the whole
00:17:03.629 --> 00:17:07.150
student population. And I think there's a different
00:17:07.150 --> 00:17:10.349
level of awareness within institutions right
00:17:10.349 --> 00:17:13.170
at the top that actions need to be taken and
00:17:13.170 --> 00:17:15.309
that changes need to be made to assessment forms,
00:17:15.309 --> 00:17:17.710
for example, in a way that we never even had
00:17:17.710 --> 00:17:20.109
a policy for machine translation at our institution.
00:17:20.390 --> 00:17:22.269
And we never had all these working groups to
00:17:22.269 --> 00:17:25.009
try and address this, even though it was radical
00:17:25.009 --> 00:17:27.599
and fundamental in the way it was changing. Yeah,
00:17:27.619 --> 00:17:29.940
I just want to come back to looking at education
00:17:29.940 --> 00:17:32.160
through the lens of friction and we'll go back
00:17:32.160 --> 00:17:34.819
to something Flau said about Confucianism. I
00:17:34.819 --> 00:17:38.180
would again push back against that characterisation
00:17:38.180 --> 00:17:40.779
of education because there's a lot of educational
00:17:40.779 --> 00:17:45.480
theory that emphasises fun and creativity and
00:17:45.480 --> 00:17:47.640
collaboration. It doesn't have to be an individual
00:17:47.640 --> 00:17:49.759
struggle. We can work collaboratively. So there
00:17:49.759 --> 00:17:52.660
are lots of different ways of thinking about
00:17:52.660 --> 00:17:57.059
education because Arguably the AI can be fun
00:17:57.059 --> 00:18:00.019
and it can be creative and there are all kinds
00:18:00.019 --> 00:18:02.839
of opportunities there to achieve that. I mean,
00:18:02.920 --> 00:18:06.380
even the friction bit, you could design tasks
00:18:06.380 --> 00:18:09.640
that challenge a user to really think about what
00:18:09.640 --> 00:18:12.559
they're prompting, asking the probing questions,
00:18:12.599 --> 00:18:14.700
you know, why have you given me this response?
00:18:14.880 --> 00:18:17.000
So there's lots of ways you can critically prompt
00:18:17.000 --> 00:18:20.420
and make and create that friction. But I suppose
00:18:20.420 --> 00:18:23.940
going back. to what you were saying, Mike, that
00:18:23.940 --> 00:18:27.240
we've been here before. I think I agree with
00:18:27.240 --> 00:18:30.700
Ayanna. Although we have been before, there is
00:18:30.700 --> 00:18:32.880
something quite significant about where we are
00:18:32.880 --> 00:18:37.799
now. I think there is a lot to be said to the
00:18:37.799 --> 00:18:41.220
fact that these technologies are not committed
00:18:41.220 --> 00:18:44.019
to the truth, that they may come up with truthful
00:18:44.019 --> 00:18:46.660
claims because the probability, the statistical
00:18:46.660 --> 00:18:51.160
probability of how they work allows for truthful
00:18:51.160 --> 00:18:54.039
things to be said. But I think the fact that
00:18:54.039 --> 00:18:56.900
what this kind of Trojan horse, as I see it,
00:18:56.940 --> 00:19:01.059
that's come into our educational system is that
00:19:01.059 --> 00:19:05.559
truth is no longer a virtue. It's no longer considered
00:19:05.559 --> 00:19:11.019
worthwhile. We know the complicated philosophical
00:19:11.019 --> 00:19:14.069
history of truth. but it's actually impacting
00:19:14.069 --> 00:19:16.289
all people's lives. And I think the machines
00:19:16.289 --> 00:19:19.769
do what Helen Beethan explains so well in her
00:19:19.769 --> 00:19:22.410
chapter in a recent publication called Critical
00:19:22.410 --> 00:19:25.849
Perspectives on Academic Writing, is Helen uses
00:19:25.849 --> 00:19:29.069
this metaphor of passing, that the idea that
00:19:29.069 --> 00:19:32.309
writing can pass as human or writing can pass
00:19:32.309 --> 00:19:35.750
as truthful without being truthful, without being
00:19:35.750 --> 00:19:38.930
committed to the truth. So I think that is quite
00:19:38.930 --> 00:19:42.990
insidious. And I think it possibly that today's
00:19:42.990 --> 00:19:46.170
conversation a little bit apart from previous
00:19:46.170 --> 00:19:52.650
discussions about threats to education. I think
00:19:52.650 --> 00:19:55.650
that goes back to what our role is, and I mean
00:19:55.650 --> 00:19:59.809
our role as education, as educators. If we are
00:19:59.809 --> 00:20:02.710
aware of what the technology does and how it
00:20:02.710 --> 00:20:06.150
operates, and yes, some of the benefits, definitely
00:20:06.150 --> 00:20:10.950
the risks, then I would say we are actually in
00:20:10.950 --> 00:20:14.109
a fairly good position to work with that or rather
00:20:14.109 --> 00:20:17.670
work against it if we want to. The idea of truth
00:20:17.670 --> 00:20:22.230
is no longer of value. Is this not actually a
00:20:22.230 --> 00:20:25.650
challenge that we need to take on actively? And
00:20:25.650 --> 00:20:27.769
that relates to what I actually said. I didn't
00:20:27.769 --> 00:20:31.069
say education needs to be hard. I said we need
00:20:31.069 --> 00:20:34.589
a challenge to learn. I'm concerned about the
00:20:34.589 --> 00:20:37.789
removal of the challenge. I think we need constructive
00:20:37.789 --> 00:20:40.640
challenges. And then we can grow and we can learn
00:20:40.640 --> 00:20:43.480
new things. And we are at a point now where we,
00:20:43.740 --> 00:20:46.220
you, me, all of us are kind of faced with a really
00:20:46.220 --> 00:20:49.119
new and challenging situation. So one thing that
00:20:49.119 --> 00:20:52.519
we can do perhaps is that we can take this into
00:20:52.519 --> 00:20:54.579
the classroom and address it with our learners
00:20:54.579 --> 00:20:57.240
and with our colleagues. And because, you know,
00:20:57.259 --> 00:20:59.220
let's not pretend that everybody has a thorough
00:20:59.220 --> 00:21:01.960
understanding of the technology. And in that
00:21:01.960 --> 00:21:04.339
sense, it is actually an opportunity to make
00:21:04.339 --> 00:21:09.329
a very risky situation. perhaps into something
00:21:09.329 --> 00:21:12.390
that could actually become constructive, I hope.
00:21:13.809 --> 00:21:15.910
Yeah, and the point of truth, again, sorry to
00:21:15.910 --> 00:21:17.490
sound like a broken record, but we've been here
00:21:17.490 --> 00:21:19.950
before. I can remember people saying, you can't
00:21:19.950 --> 00:21:22.789
use Wikipedia. Why not? Because you can't trust
00:21:22.789 --> 00:21:25.789
it. It's not the truth. When the internet first
00:21:25.789 --> 00:21:29.450
came out or expanded in a big way, people were
00:21:29.450 --> 00:21:32.230
saying there was a big thing about you can't
00:21:32.230 --> 00:21:36.599
believe what you read online. So if we are aware
00:21:36.599 --> 00:21:40.180
that what we are learning or what ChatGPT is
00:21:40.180 --> 00:21:41.640
telling us, I'm very conscious of the kind of
00:21:41.640 --> 00:21:44.779
the verbs I'm using, but what ChatGPT is suggesting
00:21:44.779 --> 00:21:47.880
to be the truth is not reliable. If we know that,
00:21:48.099 --> 00:21:50.779
then we know to fact check. And I don't think
00:21:50.779 --> 00:21:55.700
that undermines a notion of truth. I think Wikipedia
00:21:55.700 --> 00:21:58.259
was never designed with the intention to not
00:21:58.259 --> 00:22:03.740
be truthful because it is an open system. The
00:22:03.740 --> 00:22:06.960
AI technology is not designed to be truthful.
00:22:07.140 --> 00:22:09.799
Truth doesn't come into it, basically, because
00:22:09.799 --> 00:22:13.839
it is a probabilistic system. And I think this
00:22:13.839 --> 00:22:16.539
brings me to another point that I think is connected
00:22:16.539 --> 00:22:19.160
to all of this and which is also connected to
00:22:19.160 --> 00:22:21.299
something you said, Klaus, about the onus is
00:22:21.299 --> 00:22:23.640
on us as educators to help students understand
00:22:23.640 --> 00:22:26.599
this. Now, again, saying that, that sounds all
00:22:26.599 --> 00:22:30.740
very reasonable, but what that hides beneath
00:22:30.740 --> 00:22:35.029
its reasonable surface is that Do we even know
00:22:35.029 --> 00:22:37.789
what these systems are? What gives us the confidence
00:22:37.789 --> 00:22:42.210
to feel that we can educate? And I think the
00:22:42.210 --> 00:22:45.829
language that's used to talk about AI, the very
00:22:45.829 --> 00:22:49.170
phrase artificial intelligence, is incredibly
00:22:49.170 --> 00:22:52.089
misleading. The whole kind of words that are
00:22:52.089 --> 00:22:56.809
used to refer to it are misleading because they
00:22:56.809 --> 00:23:00.150
don't actually capture what this technology is.
00:23:00.309 --> 00:23:02.630
Now, one of the acronyms that I've started to
00:23:02.630 --> 00:23:06.029
use is the Salami acronym, which basically stands
00:23:06.029 --> 00:23:09.589
for systemic approaches to learning algorithms
00:23:09.589 --> 00:23:13.329
with machine inference. So calling it artificial
00:23:13.329 --> 00:23:16.470
intelligence gives the impression that it is
00:23:16.470 --> 00:23:21.049
coming up with intelligent content. It gives
00:23:21.049 --> 00:23:25.269
the impression that intelligence is one thing,
00:23:25.450 --> 00:23:28.109
that there is only one metric for measuring intelligence,
00:23:28.269 --> 00:23:31.190
and that it can therefore be artificial, a bit
00:23:31.190 --> 00:23:34.170
like an artificial heart. Now, a heart has a
00:23:34.170 --> 00:23:36.730
very specific function. There's input and output,
00:23:36.869 --> 00:23:38.750
and that's basically how it works. And that's
00:23:38.750 --> 00:23:41.950
why artificial heart can be artificial hearts.
00:23:42.450 --> 00:23:44.609
But artificial intelligence is not the same as
00:23:44.609 --> 00:23:47.750
a heart. And I think we have a problem with language,
00:23:47.950 --> 00:23:52.180
and we have a problem with what that language
00:23:52.180 --> 00:23:58.480
is signifying and how that positions us in relation
00:23:58.480 --> 00:24:02.240
to the technology. I think we've certainly got
00:24:02.240 --> 00:24:04.059
a problem with language and as I said earlier,
00:24:04.119 --> 00:24:06.960
you know, you hear people talking about the AI
00:24:06.960 --> 00:24:09.940
knows or the AI understands and I think we all
00:24:09.940 --> 00:24:12.200
would all agree that it is probabilistic. It
00:24:12.200 --> 00:24:14.859
doesn't have knowledge in the same way that we
00:24:14.859 --> 00:24:19.099
have knowledge as people. I think This comes
00:24:19.099 --> 00:24:20.980
back to the discussion we were having right at
00:24:20.980 --> 00:24:24.299
the beginning about these tools not being designed
00:24:24.299 --> 00:24:26.940
specifically for learning and that they shouldn't
00:24:26.940 --> 00:24:29.900
be leading the narrative on education. While
00:24:29.900 --> 00:24:33.019
we may not have the greatest level of understanding
00:24:33.019 --> 00:24:35.960
to be able to discuss these things fully with
00:24:35.960 --> 00:24:38.599
our students, what we do have an understanding
00:24:38.599 --> 00:24:40.980
on is the learning process and what we're teaching
00:24:40.980 --> 00:24:43.900
our students and that should be what we're centering,
00:24:44.180 --> 00:24:48.240
not these tools. Being able to discuss with students
00:24:48.240 --> 00:24:51.099
what we hope to share, what we hope they learn.
00:24:51.599 --> 00:24:53.779
Why we are teaching what we're teaching, why
00:24:53.779 --> 00:24:55.940
we're assessing in the way we're assessing is
00:24:55.940 --> 00:24:58.799
a part to me of helping them understand whether
00:24:58.799 --> 00:25:02.319
it's worth outsourcing or sitting with the trouble
00:25:02.319 --> 00:25:06.299
because that's the key purpose really in that
00:25:06.299 --> 00:25:08.779
particular teaching. And that may or may not
00:25:08.779 --> 00:25:11.359
then therefore include generative AI. It may
00:25:11.359 --> 00:25:13.599
include discussion about how that might diminish
00:25:13.599 --> 00:25:16.019
learning. It may include discussion about the
00:25:16.019 --> 00:25:18.779
reasons why students might make individual personal
00:25:18.779 --> 00:25:21.380
choices for ethical reasons to not use these
00:25:21.380 --> 00:25:24.220
tools or why they may feel for other reasons
00:25:24.220 --> 00:25:26.000
that it supports them in their learning. But
00:25:26.000 --> 00:25:29.779
the starting point is the learning, not these
00:25:29.779 --> 00:25:31.839
tools. And I think we need to recognise that
00:25:31.839 --> 00:25:35.240
that's where our expertise lies. Yeah, there
00:25:35.240 --> 00:25:38.940
are lots of scholars. at the moment that would
00:25:38.940 --> 00:25:41.220
absolutely argue that. And the way that I think
00:25:41.220 --> 00:25:44.140
that they would put it would be in terms of centering
00:25:44.140 --> 00:25:47.099
values rather than the technology. So the value
00:25:47.099 --> 00:25:49.920
of honesty, the value of truth, the value of
00:25:49.920 --> 00:25:52.619
transparency, the value of collegiality, the
00:25:52.619 --> 00:25:57.380
value of knowledge. And once you flip the framing,
00:25:57.519 --> 00:26:00.660
you stop focusing on the technology and all of
00:26:00.660 --> 00:26:04.019
its eugenic implications. It's all in the literatures
00:26:04.019 --> 00:26:07.210
on generative AI. They don't have to be my values
00:26:07.210 --> 00:26:09.829
as a teacher, but they are values and basically
00:26:09.829 --> 00:26:13.650
we all have values, whichever side of the debate
00:26:13.650 --> 00:26:18.849
we fall on. We are all motivated and we see the
00:26:18.849 --> 00:26:21.450
world through the values that we have, whether
00:26:21.450 --> 00:26:24.609
it's objectivity, whether it's neutrality, whether
00:26:24.609 --> 00:26:28.349
it's partisan politics, whenever that is. But
00:26:28.349 --> 00:26:31.750
by centering those values or foregrounding them,
00:26:31.849 --> 00:26:34.250
at least that gives everyone a kind of equal
00:26:34.250 --> 00:26:36.769
terrain around which to discuss, and then that's
00:26:36.769 --> 00:26:39.210
where the criticality can come in, and then choices
00:26:39.210 --> 00:26:48.490
can be maybe more within our reach. To a point
00:26:48.490 --> 00:26:52.569
that I hope fits, since you said values a lot,
00:26:52.730 --> 00:26:57.690
I think There is a systemic issue as in the education
00:26:57.690 --> 00:27:00.990
system or the societal system even. When I hear
00:27:00.990 --> 00:27:03.009
people talk about education, higher education,
00:27:03.109 --> 00:27:05.269
why would I get a degree and all that? All of
00:27:05.269 --> 00:27:07.710
this comes down to something countable. It comes
00:27:07.710 --> 00:27:10.150
down to how much money will I make? It comes
00:27:10.150 --> 00:27:12.470
down to employability figures. You put a number
00:27:12.470 --> 00:27:15.549
on it, it doesn't matter. I think that has lost
00:27:15.549 --> 00:27:18.230
sight of the core mission of education in general.
00:27:18.529 --> 00:27:20.990
We all need to pay the rent and all that. But
00:27:20.990 --> 00:27:24.069
I think the actual mission of education is or
00:27:24.069 --> 00:27:27.769
should be or has been at least or was to form
00:27:27.769 --> 00:27:32.589
responsible, growing individuals, citizens and
00:27:32.589 --> 00:27:35.750
not how much money will I make in the end. Now,
00:27:35.750 --> 00:27:38.170
if we have a technology and how quickly can I
00:27:38.170 --> 00:27:40.269
get there? In a time where we're talking about
00:27:40.269 --> 00:27:43.130
reduced degrees, short two year bachelor's degrees
00:27:43.130 --> 00:27:45.289
and all that. So everything has to be faster,
00:27:45.329 --> 00:27:47.789
more and so on. And I think the technology feeds
00:27:47.789 --> 00:27:50.859
right into that. any technology actually, not
00:27:50.859 --> 00:27:53.619
only AI. And this is to me a systemic problem
00:27:53.619 --> 00:27:56.619
when what we are saying now and what the institutions
00:27:56.619 --> 00:28:00.140
sometimes value or society values may actually
00:28:00.140 --> 00:28:02.660
not align very well. So that just compounds the
00:28:02.660 --> 00:28:05.900
issue. And I think it also kind of obscures the
00:28:05.900 --> 00:28:08.440
fact that not everybody is motivated by money.
00:28:08.539 --> 00:28:10.460
I mean, I can only speak for myself. But if I
00:28:10.460 --> 00:28:13.220
was motivated by money, I could have gone into
00:28:13.220 --> 00:28:16.079
the finance sector if I wanted to make money.
00:28:16.119 --> 00:28:19.119
But now I've ended up in education. where I'm
00:28:19.119 --> 00:28:22.039
not making money and I'm working overtime. But
00:28:22.039 --> 00:28:24.180
there is a sense that this narrative that we
00:28:24.180 --> 00:28:27.859
only do things because of financial motivation
00:28:27.859 --> 00:28:33.400
is actually wrong. It's not evidence based. A
00:28:33.400 --> 00:28:35.880
closing question to draw all of this together.
00:28:36.640 --> 00:28:38.299
So one of the issues that has really emerged
00:28:38.299 --> 00:28:42.599
from this is the idea that historically, somehow
00:28:42.599 --> 00:28:44.980
in the face of all these different advances in
00:28:44.980 --> 00:28:47.730
technology, even things like gamification. 15,
00:28:47.730 --> 00:28:49.650
20 years ago were going to be the new thing.
00:28:50.190 --> 00:28:54.329
We've always managed to absorb these systems
00:28:54.329 --> 00:28:58.190
into our practice as educators. Can we do this
00:28:58.190 --> 00:29:02.329
with generative AI? Or is this something like
00:29:02.329 --> 00:29:05.910
I read recently, someone said that this is the
00:29:05.910 --> 00:29:09.190
equivalent of the tide of change that came when
00:29:09.190 --> 00:29:12.769
King Henry dissolved the monasteries. Are we
00:29:12.769 --> 00:29:15.779
really heading towards that point? I know this
00:29:15.779 --> 00:29:19.000
is going to sound a little bit sort of nerdy
00:29:19.000 --> 00:29:22.019
bloke who lives by himself, but I've been playing
00:29:22.019 --> 00:29:26.220
with some vibe coding recently and it's so easy,
00:29:26.539 --> 00:29:29.819
so easy to set up a website now that can play
00:29:29.819 --> 00:29:32.980
a game with a student or can give feedback to
00:29:32.980 --> 00:29:35.160
a student or focus on a certain language point.
00:29:35.980 --> 00:29:40.740
So there's ways of doing things that really can.
00:29:41.049 --> 00:29:45.009
build on our pedagogy and do things in a way
00:29:45.009 --> 00:29:48.690
that we could have done before if we paid a web
00:29:48.690 --> 00:29:52.029
designer. But there's so much potential for increasing
00:29:52.029 --> 00:29:55.849
and improving and enhancing what we do. I think
00:29:55.849 --> 00:30:00.349
there is that. And if we only focus on all the
00:30:00.349 --> 00:30:02.130
things we said about truth, all the things we
00:30:02.130 --> 00:30:04.569
said about reliability and so on and so forth,
00:30:04.829 --> 00:30:07.210
then we're actually missing a part of the discussion.
00:30:07.630 --> 00:30:09.789
I think that is important to bear in mind. What
00:30:09.789 --> 00:30:12.750
are then the opportunities of how this could
00:30:12.750 --> 00:30:15.450
actually make a positive difference? And I'm
00:30:15.450 --> 00:30:17.549
not talking about this as you put it, Julia,
00:30:17.829 --> 00:30:20.470
in technical terms as a generator. I'm talking
00:30:20.470 --> 00:30:23.950
about actually creating something that helps
00:30:23.950 --> 00:30:27.589
people learn and that is possible as well. It
00:30:27.589 --> 00:30:30.230
is not only about spitting out text and more
00:30:30.230 --> 00:30:33.190
text. It is about things that I'm doing something
00:30:33.190 --> 00:30:36.799
similar and I'm also kind of trying to control
00:30:36.799 --> 00:30:38.839
the technology a little bit through work with
00:30:38.839 --> 00:30:41.720
AI agents that you can set up quite easily and
00:30:41.720 --> 00:30:43.599
it gives you a little bit of influence on how
00:30:43.599 --> 00:30:46.019
they operate and what they do. And if we manage
00:30:46.019 --> 00:30:48.619
to go into that direction as well as taking a
00:30:48.619 --> 00:30:51.079
critical stance to all of this, then we may come
00:30:51.079 --> 00:30:53.180
out with something that is actually helpful for
00:30:53.180 --> 00:30:55.720
our students and for our practice. Something
00:30:55.720 --> 00:30:58.259
that we couldn't do before. Yeah, I suppose my
00:30:58.259 --> 00:31:02.000
concern is that this kind of over determining
00:31:02.000 --> 00:31:04.980
the kind of optimism around it. can downplay
00:31:04.980 --> 00:31:07.759
the real harms that are coming out of it. I mean,
00:31:07.799 --> 00:31:11.700
there are real harms in terms of how these machines
00:31:11.700 --> 00:31:14.779
are trained, the kind of slave labor that they
00:31:14.779 --> 00:31:18.640
rely on that is invisible, and the kind of post
00:31:18.640 --> 00:31:21.359
-traumatic stress that's created, that's endured
00:31:21.359 --> 00:31:25.220
by people, non -Western countries that are being
00:31:25.220 --> 00:31:29.819
paid a pittance to remove content that is harmful.
00:31:30.339 --> 00:31:33.940
I think there's a huge environmental impact that
00:31:33.940 --> 00:31:37.359
is really, really serious and is unregulated
00:31:37.359 --> 00:31:40.039
because these big tech companies are not paying
00:31:40.039 --> 00:31:43.079
their taxes. They're not declaring how much energy
00:31:43.079 --> 00:31:47.000
is actually being used. The AI is being used
00:31:47.000 --> 00:31:51.420
for warfare. So there is a really, really big,
00:31:51.420 --> 00:31:55.940
cool elephant in the room. And whilst I agree,
00:31:56.059 --> 00:32:00.240
there are always affordances to everything, opportunities
00:32:00.240 --> 00:32:02.700
and challenges. There's always an argument to
00:32:02.700 --> 00:32:05.339
say, well, it could work cars, you know, the
00:32:05.339 --> 00:32:07.599
technology of cars, we know they pollute, we
00:32:07.599 --> 00:32:09.640
know they damage the environment, but we also
00:32:09.640 --> 00:32:13.940
know that they are very handy. So yes, of course,
00:32:14.099 --> 00:32:19.400
but I think not enough is discussed at the level
00:32:19.400 --> 00:32:23.079
of the sort of individual interactions that we
00:32:23.079 --> 00:32:25.299
have with these machines that not enough is said
00:32:25.299 --> 00:32:29.119
about. why this technology exists in the way
00:32:29.119 --> 00:32:32.660
it does and the harm that it does. I mean, the
00:32:32.660 --> 00:32:35.059
technology exists like any technology because
00:32:35.059 --> 00:32:37.920
it can. You're right about the environmental
00:32:37.920 --> 00:32:40.440
impacts. You're right about some of the social
00:32:40.440 --> 00:32:44.099
impacts. You're right about nobody likes war,
00:32:44.539 --> 00:32:47.180
but war, there has always been a technological
00:32:47.180 --> 00:32:50.660
competition more than most other forms of human
00:32:50.660 --> 00:32:56.259
endeavours. However, refusing to engage with
00:32:56.259 --> 00:32:58.720
it at an educational level is going to solve
00:32:58.720 --> 00:33:02.480
none of those problems. If we engage with it
00:33:02.480 --> 00:33:08.400
and embed into our education the ethics, the
00:33:08.400 --> 00:33:12.299
environmental aspects of it, the problems and
00:33:12.299 --> 00:33:16.000
everything else and we deal with it as part of
00:33:16.000 --> 00:33:19.940
our world that has come in recently through whatever
00:33:19.940 --> 00:33:22.019
ideological lens we want. You can deal with it
00:33:22.019 --> 00:33:25.839
as a marvelous capitalist innovation and creative
00:33:25.839 --> 00:33:27.619
destruction, all of that kind of stuff. You can
00:33:27.619 --> 00:33:29.900
deal with it as an environmental threat, whatever.
00:33:30.259 --> 00:33:33.819
But if we deal with that and communicate and
00:33:33.819 --> 00:33:38.519
encourage our students to think about it in a
00:33:38.519 --> 00:33:41.519
sensible, measured, and evidence -based way,
00:33:42.079 --> 00:33:44.339
that, I believe, is the way that we will help
00:33:44.339 --> 00:33:46.839
our students, because it's going to be their
00:33:46.839 --> 00:33:48.599
world, it's not going to be our world for much
00:33:48.599 --> 00:33:54.900
longer, thrive in an AI Should I say AI ubiquitous
00:33:54.900 --> 00:33:57.319
world? Yeah. I think I'd go back to something
00:33:57.319 --> 00:33:59.700
I said at the beginning. If the whole system
00:33:59.700 --> 00:34:04.779
is set up to achieve quick wins, quick fixes,
00:34:05.220 --> 00:34:09.559
easy, smooth rides, efficiency is people go to
00:34:09.559 --> 00:34:12.539
universities to get a degree, not to learn. If
00:34:12.539 --> 00:34:15.119
everything you've said is set within a context
00:34:15.119 --> 00:34:20.159
that in a sense is not conducive to thinking.
00:34:21.119 --> 00:34:24.940
about it in this nuanced and critical way. There
00:34:24.940 --> 00:34:27.400
is hope and there is possibility in thinking
00:34:27.400 --> 00:34:29.519
that actually, no, we can think about all of
00:34:29.519 --> 00:34:32.559
this afresh. We can break down systems and we
00:34:32.559 --> 00:34:36.480
can reimagine the future that we would like to
00:34:36.480 --> 00:34:40.380
inhabit. Okay, I was just jumping in with one
00:34:40.380 --> 00:34:42.480
last thing to really come back to what Klaus
00:34:42.480 --> 00:34:45.719
was saying about numbers and the issue around
00:34:45.719 --> 00:34:49.340
an education being to earn more and how we've
00:34:49.340 --> 00:34:53.389
lost. in some respects because a degree has become
00:34:53.389 --> 00:34:56.010
much more of a usual thing. There's a much larger
00:34:56.010 --> 00:34:59.429
number of people taking first degrees. The value
00:34:59.429 --> 00:35:01.929
of a degree has been this sort of specialist,
00:35:02.530 --> 00:35:06.289
privileged, elitist credit that would for certain
00:35:06.289 --> 00:35:08.730
get you a jump at some point so you could enjoy
00:35:08.730 --> 00:35:11.329
and focus on the learning. That has shifted and
00:35:11.329 --> 00:35:13.429
we've become much more of a student as consumer.
00:35:13.650 --> 00:35:14.909
I think we haven't talked about that enough.
00:35:15.159 --> 00:35:17.420
today and I know we're running out of time but
00:35:17.420 --> 00:35:19.599
I think that's a huge impact on this idea of
00:35:19.599 --> 00:35:22.000
valuing the product to the end result over the
00:35:22.000 --> 00:35:24.460
process and that's got to be the context in which
00:35:24.460 --> 00:35:28.079
we understand use of or overuse of generative
00:35:28.079 --> 00:35:32.460
AI. So to me it's about recognising that some
00:35:32.460 --> 00:35:35.280
things have changed in terms of higher education
00:35:35.280 --> 00:35:38.480
and its socio -political economic context but
00:35:38.480 --> 00:35:42.179
that there are also very many good outcomes of
00:35:42.179 --> 00:35:44.659
that in that we have more people accessing higher
00:35:44.659 --> 00:35:47.340
education. To put it down to a sort of massification
00:35:47.340 --> 00:35:50.179
in a negative sense is problematic in and of
00:35:50.179 --> 00:35:52.500
itself. So we need to then think about how do
00:35:52.500 --> 00:35:55.219
we bring back the process and the learning and
00:35:55.219 --> 00:35:58.579
to me that's again it's starting about from censoring
00:35:58.579 --> 00:36:02.260
the learning and including all of the areas that
00:36:02.260 --> 00:36:05.239
Julia has talked about. And actually I think
00:36:05.239 --> 00:36:07.179
a lot of the ways in which we set up that learning
00:36:07.179 --> 00:36:10.260
impact on how students engage with generative
00:36:10.260 --> 00:36:12.400
AI. And it's not about just what we do as individual
00:36:12.400 --> 00:36:14.900
educators, it's about us pushing and pushing
00:36:14.900 --> 00:36:18.780
and pushing on the system in which we work to
00:36:18.780 --> 00:36:21.300
value that process. So things, basic things like
00:36:21.300 --> 00:36:22.800
making sure that students actually have time
00:36:22.800 --> 00:36:26.219
to engage with assessments properly, to put themselves
00:36:26.219 --> 00:36:28.940
into it and for us to mark them properly and
00:36:28.940 --> 00:36:31.500
to value students bringing themselves to that
00:36:31.500 --> 00:36:34.550
assessment. impact on how students engage with
00:36:34.550 --> 00:36:36.349
generative AI and how they engage with their
00:36:36.349 --> 00:36:38.349
learning process. It's not really about generative
00:36:38.349 --> 00:36:40.150
AI, is it? It's about all of those things about
00:36:40.150 --> 00:36:42.349
education that have always been the case, but
00:36:42.349 --> 00:36:45.730
they've become ever so much more important now,
00:36:45.769 --> 00:36:49.469
I think. And that's the work for us to do, yes,
00:36:49.469 --> 00:36:52.110
in our own practice with our students, but also
00:36:52.110 --> 00:36:56.329
in pushing systemically. OK, thank you, everyone.
00:36:56.610 --> 00:36:59.389
I think that's going to bring this episode towards
00:36:59.389 --> 00:37:03.170
the end. And now, before we wrap up today, thank
00:37:03.170 --> 00:37:05.210
you so much for listening to this episode. We
00:37:05.210 --> 00:37:07.170
hope you've enjoyed the conversation and found
00:37:07.170 --> 00:37:10.530
it useful. As an added bonus, if today's discussion
00:37:10.530 --> 00:37:12.949
has sparked your interest, many of the themes
00:37:12.949 --> 00:37:15.230
we have explored are examined in greater depth
00:37:15.230 --> 00:37:18.469
in the recently published book, Critical Perspectives
00:37:18.469 --> 00:37:21.269
on Academic Writing, The Politics and Practices
00:37:21.269 --> 00:37:24.369
of Writing Development in UK Higher Education,
00:37:25.010 --> 00:37:28.010
edited by Ayanna, one of our guests today. along
00:37:28.010 --> 00:37:31.469
with David Manos and published by UCL Press.
00:37:31.929 --> 00:37:34.550
It brings together a range of perspectives on
00:37:34.550 --> 00:37:37.570
academic writing, including some of the challenges
00:37:37.570 --> 00:37:39.969
and opportunities presented by generative AI
00:37:39.969 --> 00:37:43.190
in higher education. And it's certainly worth
00:37:43.190 --> 00:37:46.750
accessing if you want more information on some
00:37:46.750 --> 00:37:48.929
of the themes that we've discussed today. Thank
00:37:48.929 --> 00:37:50.869
you very much for listening and see you next
00:37:50.869 --> 00:37:57.090
time. Thank you. We really want to hear from
00:37:57.090 --> 00:37:59.849
you as well So if you'd like to share a comment
00:37:59.849 --> 00:38:02.110
suggestion or an insight into something that
00:38:02.110 --> 00:38:04.670
we've talked about on one of the episodes Please
00:38:04.670 --> 00:38:07.510
send us an email or voice note at podcast at
00:38:07.510 --> 00:38:10.849
barley org But you can also get in touch through
00:38:10.849 --> 00:38:14.469
our YouTube channel at all things EAP Thank you
00:38:14.469 --> 00:38:16.650
again for listening and we hope that you can
00:38:16.650 --> 00:38:19.869
join us for the next episode. Take care. Bye.
00:38:19.869 --> 00:38:20.170
Bye