May 20, 2024
With the public release of large language models like Chat GPT putting Artificial Intelligence (AI) firmly on our radar, this episode explores what benefits this technology might hold for statistics and analysis, as well as policymaking and public services.
Joining host, Miles Fletcher, to discuss the groundbreaking work being done in this area by the Office for National Statistics (ONS) and across the wider UK Government scene are: Osama Rahman, Director of the ONS Data Science Campus; Richard Campbell, Head of Reproducible Data Science and Analysis; and Sam Rose, Deputy Director of Advanced Analytics and Head of Data Science and AI at the Department for Transport.
Transcript
MILES FLETCHER
Welcome again to Statistically Speaking, the official podcast of
the UK’s Office for National Statistics. I'm Miles Fletcher and, if
you've been a regular listener to these podcasts, you'll have heard
plenty of the natural intelligence displayed by my ONS colleagues.
This time though, we're looking into the artificial stuff. We'll
discuss the work being done by the ONS to take advantage of this
great technological leap forward; what's going on with AI across
the wider UK Government scene; and also talk about the importance
of making sure every use of AI is carried out safely and
responsibly. Guiding us through that are my ONS colleagues - with
some of the most impressive job titles we've had to date - Osama
Rahman is Director of the Data Science Campus. Richard Campbell is
Head of Reproducible Data Science and Analysis. And completing our
lineup, Sam Rose, Deputy Director of Advanced Analytics and head of
data science and AI at the Department for Transport. Welcome to you
all. Osama let's kick off then with some clarity on this AI thing.
It's become the big phrase of our time now of course but when it
comes to artificial intelligence and public data, what precisely
are we talking about?
OSAMA
RAHMAN
So
artificial intelligence quite simply is the simulation of human
intelligence processes by computing systems, and the simulation is
the important bit, I think. Actually, people talk about data
science, and they talk about machine learning - there's no
clear-cut boundaries between these things, and there's a lot of
overlap. So, you think about data science. It's the study of data
to extract meaningful insights. It's multidisciplinary – maths,
stats, computer programming, domain expertise, and you analyse
large amounts of data to ask and answer questions. And then you
think about machine learning. So that focuses on the development of
computer algorithms that improve automatically through experience
and by the use of data. So, in other words, machine learning
enables computers to learn from data and make decisions or
predictions without explicitly being programmed to do so. So, if
you think about some of the stuff we do at the ONS, it's very
important to be able to take a job and match it to an industrial
classification - so that was a manually intensive process and now
we use a lot of machine learning to guide that. So, machine
learning is essentially a form of AI.
MILES
FLETCHER
So
is it fair to say then that the reason, or one of the main reasons,
people are talking so much about AI now is because of the public
release of these large language models? The chat bots if you like,
to simpletons like me, the ChatGPT’s and so forth. You know, they
seem like glorified search engines or Oracles - you ask them a
question and they tell you everything you need to know.
OSAMA
RAHMAN
So
that's a form of AI and the one everyone's interested in. But it's
not the only form – like I said machine learning, some other
applications in data science, where we try in government, you know,
in trying to detect fraud and error. So, it's all interlinked.
MILES
FLETCHER
When
the ONS asked people recently for one of its own surveys, about how
aware the public are about artificial intelligence, 42% of people
said they used it in their home recently. What sort of things would
people be using it for in the home? What are these everyday
applications of AI and I mean, is this artificial intelligence
strictly speaking?
OSAMA
RAHMAN
If
you use Spotify, or Amazon music or YouTube music, they get data on
what music you listen to, and they match that with people who've
been listening to similar music, and they make recommendations for
you. And that's one of the ways people find out about new music or
new movies if you use Netflix, so that's one pretty basic
application, that I think a lot of people are using in the
home.
MILES
FLETCHER
And
when asked about what areas of AI they'd like to know more about,
more than four in 10 adults reported that they'd like to know
better how to judge the accuracy of information. I guess this is
where the ONS might come in. Rich then, if I could just ask you to
explain what we've been up to, what the Data Science Campus has
been up to, to actually bring the power of artificial intelligence
to our statistics.
RICHARD
CAMPBELL
Thanks
Miles. Yeah, a few things that ONS has been doing in this very
broad sphere of artificial intelligence, and it's really in that
overlap area that Osama mentioned with data science, so I'd pick
out a few sorts of general areas there. So, one is automation. You
know, we're always keen to look at how we can automate processes
and make them more efficient. It frees up the time of our analysts
to conduct more work. It means that we are more cost effective. It
means that our statistics have better quality. It's something we've
done for years but AI offers some new opportunities do that. The
other area which Osama touched on is the use of large language
models, you know, we can get into the complexities of data. We can
get much more out of data; we can complete tasks that would have
been too complex or too time consuming for real data scientists.
And this is good news, actually, because it frees up the data
scientists to add real valuable human insights. Some of the places
we've been using this. So, my team for example, which is called
reproducible data science and analysis, and we use data science and
engineering skills to develop computer systems to produce
statistics where the data is a bit big, or what I tend to call a
bit messy or a bit complex for our traditional computer systems. We
use AI here through automation, as I mentioned, you know, really
making sure that we're making systems as efficient and high quality
as possible. Another thing we're interested in doing here is quite
often we’re doing something called re-platforming systems. So, this
is where we take a system that's been used to produce our
statistics for years and years and look to move it on to new
technology. Now we're exploring with Osama's team the potential for
AI to do a lot of the grunt work for us there to sort of go in and
say, right, what is going on in this system? How is it working, how
we can improve it? One other thing I'll mention, if Osama doesn't
mind me treading on the territory of his team, is the Stats Chat
function that we've used on the ONS website. So, this is using AI
to enable a far more intelligent interrogation of the vast range of
statistics that we've got, so it no longer requires people to be
really knowledgeable about our statistics. It enables them to ask
quite open questions and to be guided to the most relevant
data.
MILES
FLETCHER
Because
at the moment, if you want to really explore a topic by getting
into the depths of the data, into the granular data, you’ve really
got to know what you're looking for haven’t you? This again is an
oracle that will come up with the answers for you and just present
them all ready for your digestion.
RICHARD
CAMPBELL
That's
right. And I tend to think of these things as a starting point,
rather than the whole answer. So, what it’s enabling you to do is
to get to the meat of the issue a lot quicker. And then you can
focus your energy as a user of our statistics in doing the analysis
that you want rather than thinking “how do I find the right
information in the first place?”
MILES
FLETCHER
Osama,
that sounds like an intriguing tool. Tell us precisely how it works
then, what data does it capture, what's in scope?
OSAMA
RAHMAN
So
the scope is publicly available documents on the ONS website. And
there's a specific reason for that. So, these AI tools, you can
have it look at the whole internet, you can have it look at subsets
of data, you can point it to specific bits of data, right? And
what's important for us is actually the work of the ONS, that
statistics we produce are quality assured and relevant. And by
providing these guardrails where you know, Stats Chat only looks at
ONS published data, we have a degree of assurance that the data
coming back to the user is likely to be of good quality and not
based on who knows what information.
MILES
FLETCHER
Because
when you use, to name one example, ChatGPT for example, the little
warning comes back saying “ChatGPT can make mistakes, consider
checking important information.” And I guess that's fundamental to
all this isn't it. These tools, as intelligent as they might be,
they're only as good - like any system - as the information that's
going in the front end.
OSAMA
RAHMAN
That's
absolutely correct, which is why we have these guardrails where,
you know, the functionality on Stats Chat is focused on published
ONS information.
MILES
FLETCHER
That
does mean that something that's offered by an organisation like the
ONS does have that sort of inbuilt potential to be trustworthy and
widely used. But of course, you might say, to have a really good
tool it's got to be drawing on masses of information from right
across the world. And it's interesting how, and you mentioned that
it's open-source data, of course, that's most available for these
tools at the moment, but you're seeing proprietary data coming in
as well. And this week, as we're recording this, the Financial
Times, for example, has announced that it's done a deal with one of
the big AI firms to put all of its content into their database. Do
you think there's scope for organisations like the ONS around the
world to collaborate on this and to provide you know, really
powerful tools for the world to exchange knowledge and data this
way?
OSAMA
RAHMAN
So
there is collaboration going on. There's collaboration, both within
government - we're not the only department looking at these sorts
of tools; there's also collaboration internationally. I think the
difference you know... our information on our website is already
publicly available. That's why it's on the net, it is a
publication. But there's a difference in situation with the FT
where, you know, a lot of the FT information is behind a
paywall.
MILES
FLETCHER
Yeah,
it has a sort of democratising tendency that this publicly
available information is being fed into these kinds of sources and
these kinds of tools. That's big picture stuff. It's all very
exciting work that's going on. But I'll come back to you Rich just
for a second. What examples practically, because I think that the
Stats Chat project is still a little way off actually being
available publicly, isn't it?
RICHARD
CAMPBELL
Yeah,
I think it is still a little way off. So, I think the key thing
that we're doing at the moment and something we've done for years,
but AI is helping is the use of automation principles. Just making
things quicker. Now in a data science context, this might be going
through very, very large data sets, looking for patterns that it
would take an analyst a huge amount of time and probably far too
much patience than they would have to find.
MILES
FLETCHER
So
for example, in future then we might find that - and this is one
issue that recurs in these podcasts - obviously about the
limitations of official statistics is they tend to lag.
This is another way
of making sure that data gets processed faster. And therefore, the
statistics are more timely, and therefore the insights they provide
are really much more actionable than perhaps they might be at the
moment.
RICHARD
CAMPBELL
Yeah,
that's spot on. There's potential in there for pace of getting the
statistics from the point that the data exists to getting it into
published statistics. There's potential there for us to be able to
combine and bring more sources together. There's also some behind
the scenes stuff that helps as well. So, for example, quite often
we are coding up the systems to produce new or improved versions of
official statistics. And we're looking at the possibility of AI
speeding up and supporting that process, perhaps for example, by
giving us an initial draft of the code. Now, why does that matter
for people in the public, you know, does anybody actually care?
Well, what it means is that we can do things quicker and more to
the point we can focus the time of our expert data scientists and
other analysts in really helping people understand the data and the
analysis that we're producing.
MILES
FLETCHER
Okay,
so lots of interesting stuff in the pipeline there. But I’d like to
bring in Sam now to talk about how AI is actually being used in
government right now. Because in your work Sam at the Department
for Transport, you've actually been working on some practical
projects that have been gaining results in the real
world.
SAM
ROSE
We
have - we've been doing loads actually, and my poor team probably
haven't had any time to sit still for the last 18 months or so. And
I think like most ministerial departments, we're doing lots and
lots of work to automate existing processes, so much like Rich has
alluded to in your space, we're looking at the things that take up
most of the time for our policy colleagues and looking at how we
can automate those. So, for example, drafting correspondence, or
automating policy consultation processes, or all of that kind of
corporate memory type stuff. Can we mine big banks of data be it
text or otherwise and summarise that information or generate new
insights that we wouldn't have been able to do previously? But I
think slightly more relevant maybe for you guys, is the stuff we're
doing on creating new datasets or improving datasets. So, a few
things. We're training a machine learning model to identify heavy
goods vehicles from Earth observation data. And that's because we
don't have a single nationally representative data set that tells
us where these heavy goods vehicles park or stop outside of
existing kind of service stations, and what we want to understand
is where are those big areas of tarmac or concrete where they're
all parking up as part of their routine journeys, so that we can
look at when we're rolling out the green infrastructure for heavy
goods vehicles, we're looking at where the important places that we
need to put that infrastructure are. And that data doesn't exist at
the moment. So we're using machine learning to generate a new
dataset that we wouldn't otherwise have.
MILES
FLETCHER
And
how widespread are these kinds of projects across government in the
UK now?
SAM
ROSE
So
I think that there are loads of different things and I wouldn't be
able to speak on behalf of everybody but I know lots of different
areas of government are looking at similar kind of automation and
productivity projects like our kind of drafting all of the
knowledge management area. I think there's things like Osama
alluded to where DEFRA for example, I think they're using Earth
observation data to assess biodiversity for example. So, there's
lots of stuff that's common between lots of government departments,
and then there's lots of stuff that's very specific to individual
departments. But all along the way there's lots of collaboration
and working together to make sure we're all learning continuously
and where we can collaborate on a single solution that we
are.
MILES
FLETCHER
I
guess one of the central public concerns about the spread of AI
once again that it will cost jobs, that it will do people out of
the means of making a living that they've become used to. And I
guess from government's point of view, it's all about doing much,
much more with the resources that we have and making government
much more effective.
SAM
ROSE
Yes,
absolutely. And it's not necessarily - and I think Rich mentioned
this earlier - it's not necessarily about doing our jobs for us.
It's about improving how we can do our jobs and being able to do
more with less, I think, so freeing up the human to do the bit that
the human really needs to do and enabling the technology to do
their very repeatable very automatable parts of the job. And
indeed, in some instances, this technology can actually do the work
better than humans. So be it identifying really complex patterns
and datasets, for example. Or a good example from us in transport
is we've trained machine learning model to be able to look at
images of electric vehicle charge point installations and be able
to identify that similar or the same image that has been submitted
more than once. Now that's estimated to have saved over 130 man
years of time, you know, that's not a task that we would have been
able to do with just humans.
MILES
FLETCHER
And
you would have to be pretty alert as a human and have a very high
boredom threshold to process all that material yourself and spot
the fraudsters.
SAM
ROSE
Yeah,
well, quite. And that's, I think, a really nice example of where
again, it's not taking our jobs, but it's enabling us to do
something that we wouldn't have been able to do previously and
improve the service that we're providing.
MILES
FLETCHER
Now,
our ability collectively, whatever sort of organisation we're
involved in, our ability to make the most of AI depends on of
course having the right skills, and Osama I guess this is where the
Data Science Campus comes in as the government's Centre of
Excellence for data science, principally, but I guess also in this
context, artificial intelligence as well. What work have you been
involved in to make sure that the supply of those skills and
knowledge is on tap for government?
OSAMA
RAHMAN
So
firstly, I would say we are a (one) centre of excellence within
government. I think you know, what's been brilliant to see since
the campus was set up has been that actually more and more
government departments have excellent data science, AI teams. Sam
leads one at DfT. There is, of course, 10DS (or 10 Data Science) at
number 10 [Downing Street]. There's a Cabinet Office team. So,
there's lots of teams that now work in this area. Some of the stuff
we've been doing is we have various training programmes that we
have run. We have senior data masterclasses so that actually,
senior leaders within government can understand better the power of
data. 10DS, Sam's area, have all been running hackathons, which
actually improve skills as well. So, it's no longer just us who are
building capability. I think it's great to see that across
government and across departments there are teams improving skills
within their departments, bringing in others from outside to work
with them. So, there's a lot going on there.
SAM
ROSE
Just
really quickly, it's important to think that skills are not just
skills of data scientists, but skills of everybody's ability to use
this kind of technology. There's a lot of work going on at the
moment looking at what we need to do both internally to government,
but also out there in all of our sectors to make sure that our
workforce has the skills it needs to be able to more rapidly kind
of adopt and be able to take advantage of all the benefits that
this technology brings to us. I mean from a very personal point of
view, and I don't really know all of the answers to this, but you
know, I'm thinking about what actually, if large language models
can help us to generate efficient code, then actually, what skills
do I need in my data scientists? If it's not writing code, is it
actually the analytical thinking and being able to understand how
to apply these kinds of technologies? So, I think it changes what
we need in the workforce that we have.
MILES
FLETCHER
Inevitably,
though, if we're talking about this kind of technology being rolled
out across government and thereby increasing the power of
government to know more about more people, then concerns obviously,
about the ethical use of data come in...
RICHARD
CAMPBELL
Maybe
if I can just come in on that one Miles. Using data safely and
responsibly - it's built into our very DNA in ONS and across
government. And our keenness to sort of learn how to do new tools
new techniques is always going to be tempered by our need to ensure
that we are responsibly using the data that's been entrusted to us.
And I think we need to sort of strike a balance here. We need to
ensure that we don't take this responsibility as an excuse to not
try and adopt new technology such as AI, but it also means we have
to do so with care and responsibility and to do it at an
appropriate pace. The key thing, I think, for me is ensuring that
we can retain control of the data that we've been entrusted with.
And so, understanding what AI is doing with that data, considering
what data we're giving access to it, what data is being processed,
and what data is being generated. And this is really at the
forefront of our minds and our collective use of this. I think our
approach - and Osama touched on this earlier - is to sort of be
novel and start with open source and non-sensitive data first, so
that will help us learn how we can effectively use it before we go
on to some of the more sensitive data that we
hold.
SAM
ROSE
We
have to have ethics and data protection at the heart of everything
we do, which then does have the tendency I think necessarily to
reduce the pace of our ability to roll things out a little bit. But
as government we do, I think have more responsibility. We can't
have those kind of oops moments that some of the big tech companies
have had when they're trying to reverse engineer the data to remove
bias and that you know, things like that that then fundamentally
undermine the output of their models. I think when you're doing a
job that affects individual people, and providing services that
affect citizens then we don't really have the luxury of getting it
wrong like that, and we have to try to make sure we get it right
first time. So, all of the things that Richard said about starting
with, you know, safer datasets and working our way up before we
deploy these models is kind of fundamental to how we're going to
learn and ensure that we're doing it safely and
securely
MILES
FLETCHER
Osama
what's your take on the ethics question?
OSAMA
RAHMAN
First
of all, I would echo everything just said. You know the Statistics
Code of Practice is an annex to the Civil Service Code, it applies
to all of us not just statisticians - I'll point that out. It is I
think, not just in the ONS, I think for analysts and data
scientists and specialists across government, this is kind of built
into their DNA. Central Digital and Data Office has put together
guidance and circulated it across government on the safe use of AI
within government. So, within government, we do take this quite
seriously. And then actually in terms of the use of some of these
techniques, I think pointing these tools at data and information
that we know is accurate is an important starting point - so having
those guardrails. If it's going to be used for decision making,
then having a human in the loop is quite important to make sure
that the use is ethical. So, there's a bunch of safety checks that
we do put in which I think allow for us to have some assurance that
the use of these tools will be safe and ethical.
RICHARD
CAMPBELL
I
think just as one additional point is you know; this isn't a new
challenge for us. It's a different flavour of a challenge that we
faced in considering new technology in the past. So, we can think
in fairly recent times the use of cloud technology to securely and
safely store data. If we go further back the use of the Internet,
go back further, again, the use of computers to hold data. And what
I think we've demonstrated time and time again, is that we do
approach these things responsibly and maturely. But we do find
opportunities to use all of them to improve the quality of
statistics and analysis and the service that we offer the
public.
MILES
FLETCHER
Looking
to the future then, and this is a very fast-moving future of
course, I'd like to get your takes on also where you see us in five
years’ time in 10 years’ time with this. I mean starting with the
Office for National Statistics – Osama and Rich particularly on
this. How will we start to see the published statistics and the big
key topics, but also the granular insights that we provide on all
kinds of areas. How will we see that changing and developing do you
think? Where are you going to put your money?
RICHARD
CAMPBELL
I
think predicting the future in this way is quite a dangerous game.
I’m thinking back to you know, if we had this podcast in the year
2000 and we asked ‘’how would the internet form part of our working
lives?’ We would have predicted something which would have been
quite different from the impact that it had. Saying all that I
think it will make a fundamental difference to the way that we
work. I see that it will be integrated in the day-to-day tasks that
we do in a similar way that we used computers to speed up and
change the way that we produced statistics. I think it will enable
our users to far better interact and engage with data and analysis.
So, it will be less of us producing a specific finalised product
for them, and more for them to be able to sort of get in ask
questions, probe and really, really interact. And I think lastly,
it will give us more potential to work and analyse data because one
thing, and I think this is really important to say, AI will give
more opportunities for analysts. It won’t take them away. It will
give them more space, more tools to work with to produce better,
more complex, more useful datasets and analysis for ONS and for its
users.
SAM
ROSE
I
was just going to add that I think it will fundamentally change the
nature of what we do. A little bit like Rich said, the sort of work
that we do will be different, but really critically, I think in a
few years’ time we won't really notice that change. I was thinking
that most people have forgotten that 10 or more years ago before
you left the house to go somewhere new, you would have consulted
your map. Whereas actually nobody, or very few people, do that
anymore. So, I think we're going to forget very quickly that lots
of what we will be doing will be AI driven.
MILES
FLETCHER
So
it's a big evolutionary step forward, if not quite a revolution. Do
you agree with that Osama?
OSAMA
RAHMAN
Absolutely,
because some of us have actually been using sort of
transformers-based models, which is what these large language
models are based on for... My team has been working with those for
at least the last eight years. But I wanted to just pick up on what
Rich just said. And it is an evolution right. And you can't
separate the tools from the data. And one of the things we're
getting now is data that is much more granular and of much higher
velocity than the data we were used to. So that allows us to look
at things at a more local level, at a more timely level. What I do
completely agree with Rich on is actually a lot of these tools and
methodologies allow the technical production of statistics to get
more efficient, which then allows you to produce more statistics at
a disaggregated level - at a regional level or local authority area
level or looking at different sub populations. It allows us to
update statistics more frequently. But then also what it allows us
to do, because it's not just about the production of the
statistics, it's about what those statistics actually tell you is
going on. And I think it allows the people we have at the ONS and
other government departments to spend more time on the real value
added which is “what does this mean?”
MILES
FLETCHER
It's
interesting if you're researching a particular topic, it must be
good to sort of evolve your methodology quickly and to refine your
processes on the run as it were to explore a particular topic. One
thing of course we need in statistics is consistency of methodology
and approach. Does that limit do you think, either of you, the
ability for statistics to get more insightful to get more germane
to issues because we have to stick to accepted methodologies to
provide that consistency over the long run?
RICHARD
CAMPBELL
I
don't think it does Miles. I mean, you're right there. There's
always a challenge for us in that, that consistency is really
important, that comparability in a time series. Equally, users do
want us to look for improvements, more detail, whether that's
granularity or whatever else. And actually, we've got a really good
successful track record of both maintaining the consistency of our
statistics, while at the same time introducing new and improved
methods. We do it with GDP. We do it with inflation, we do it with
population, it's something that we do time and time again. And,
actually, I think automation AI offers up some really exciting
opportunities here in terms of methods that can be applied. There's
actually an element of it, which will help us in the understanding
and documentation and consistent application of the methods as
well. It’s perhaps one of the less – if you don’t mind me using the
word - “sexy” applications of AI but using it to ensure that our
documentation is absolutely spot on and done quickly. To ensure
that we are applying methods really quickly and consistently. I
think AI offers us potential to do that even
better.
MILES
FLETCHER
“Sexy”
in the particular way that we refer to progress in data
science.
RICHARD CAMPBELL
Yes, quite.
OSAMA
RAHMAN
Can
I just come in on this? And it's possibly worth using a specific
example of putting out statistics on prices. In the old days you’d
basically have people going out into the field, and that's where
you'd find a basket of goods, and using pen and paper would collect
prices. Now where a lot of national statistical organisations are
going to is actually getting scanner data, because most things when
you pay for them nowadays in many parts of the world, it's scanned
first, electronically rather than rung up through a cash register
of some sort. So, scanner data provides a lot of information about
what is being purchased, and at what price it's being sold at
various retail outlets. And so, you have this data which again is
much more granular and has much higher velocity then price data you
can collect through surveys, and you know, how you integrate that
into the production of pricing statistics and other economic
statistics is really, you know, a really interesting question and
work that a lot of national statistical organisations are working
on. So, there's still the basic methodology remains the same. It's
you know, kind of defined a basket of goods, but you expand the
scale of the basket, we'll get prices on at what each of the
elements are those baskets are being sold at, and then produce a
price measure an inflation measure, right. But these tools and the
increasing quantity of data allow us to do that. But you know, the
basic methodology is kind of the same, but actually the increase in
this data allows us to do that in kind of a different way. It's an
evolution.
MILES
FLETCHER
It
does all suggest though, that perhaps the survey might finally be
replaced - the big social surveys that the ONS runs. Do you think
that the surveys days are numbered, therefore because of
AI?
RICHARD CAMPBELL
No.
OSAMA RAHMAN
No.
[LAUGHTER]
MILES
FLETCHER
A resounding no.
RICHARD CAMPBELL
That
was a resounding no, and it's not a pre-rehearsed one. And maybe
I'll just take us back Miles. So, if we went back about the best
part of 10 years, everyone was talking about big data. You know,
the days of a survey was gone. What we needed was these big,
complex, sometimes quite messy data sources that were collected for
a variety of other reasons, and that we could utilise those to sort
of answer all of the statistical questions that we had. Now, what
we found out actually is that yes, these data sources can give us a
lot of potential; data science is helping us make the most of them;
AI is helping us make even more from them. What we also learned
though, is that they work best when they're complementing the
surveys, rather than trying to replace. Think of it a bit as horses
for courses. Actually, though, I want to give an example of where
AI might be able to help us improve the response rates on surveys.
So, AI might be able to help respondents navigate through some of
the surveys, helping them understand what it is that they're being
asked. Helping them answer a bit more efficiently. So that might
actually remove a barrier that some people, some businesses have to
respond to surveys. So, you never know we might see a bit of an
uptick in response rates with a bit of AI’s help.
OSAMA
RAHMAN
And
I think the other thing I would add is what surveys are
particularly good at is getting information on the extremes of the
distribution. It's great if you think everything's going to be
generated through digital footprints data and online services, but
actually not everyone... some people... apparently dumb phones are
coming back into fashion. Or there are groups that you know, for
whatever reason, are not picked up in other forms of data. And
actually, surveys are really important for accessing, getting
information about hard to reach groups at the end of the
distribution.
MILES
FLETCHER
I
think that’s kind of reassuring that for all the promise of AI in
this brave new world, that we hope won't be a dystopian future, but
whether it will deliver all those things that we've been talking
about in terms of better insights, faster statistics, and all that.
It's still good to hear though, isn't it, that there is no
substitution from speaking to real human beings
directly.?
OSAMA
RAHMAN
I
agree entirely.
MILES
FLETCHER
Well,
that's it for another episode of Statistically Speaking and, in
summary, I suppose the use of AI feels like a natural evolution
with a number of potential benefits, and potentially huge benefits,
but with its adoption we need as always to be thoughtful and
ethical. So, thanks to all our guests: Sam Rose, Osama Rahman, and
Rich Campbell, and of course, thanks to you as always for
listening. You can subscribe to future episodes of this podcast on
Spotify, Apple podcasts and all the other major podcast platforms.
You can also follow us on X - formerly known as Twitter - via the
@ONSfocus feed. I’m Miles Fletcher and from myself and producer
Steve Milne. Until next time, goodbye.
ENDS