January 17, 2024 • 33 mins
In the first episode of our generative AI series, our panel of experts discusses the impact and potential of this emerging technology. They delve into the rise of Gen AI, the crucial role that data plays, and the insights gained by RSM UK. Join host Ben Bilsland, media and technology partner, along with guests Sarah Belsham, data analytics and insights partner, and Chris Knowles, chief digital officer from RSM UK, for an engaging conversation.
For more insights follow the link https://bit.ly/3O6KVM2
And follow us on social:
LinkedIn - https://bit.ly/3Ab7abT
Twitter - http://bit.ly/1qILii3
Instagram - http://bit.ly/2W60CWm
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Hello and welcome to The Loopwith ChatGPT recently celebrating
its first birthday.
We're going to be discussinggenerative artificial intelligence.
So joining me,our RSM chief digital officer, Chris
Knowles anddata analytics and insights partner,
Sarah Belsham.
So, Sarah, Chris, welcome.
Thank you. Great to be here.
(00:20):
Thank you.
So nobody had heard of generativeartificial intelligence until
ChatGPT.
So Sarah, whyhas Gen I taken on so much momentum
compared to all the other AI toolsthat have been out there for a
number of years?
Yeah, it's an interesting point Ben,because obviously AI isn't
a new technology, not by any means.
It's been around for, you know, 50plus years.
(00:42):
Um, but I think the traditional AIthat, um, people may
be aware of requiresquite a lot of extensive compute
and processing power, largevolumes of data, um, and
actually quite a lot of veryspecialist skills around sort
of statistical analysis of data.
So it's not for the masses byany stretch.
(01:03):
Whereas ChatGPT launchedand offered a very different flavour
of AI, something much moreaccessible to people and something
more akin to what people are usedto in their day to day lives, more
intuitive and not necessarilyrequiring the level of technical
skills that traditional AI requires.
Chris, do you do you get excitedabout Gen AI year on from ChatGPT?
(01:25):
I am excited, and I think, as Sarahsaid, it's, uh, you know, it's
building on a technology that's beenaround for a while, but it's it
reached a tipping point, didn't it,where the usability, the quality of
the natural language output becamesuch that suddenly it was
something that OpenAI, inconjunction with Microsoft, as we
know, was confident enough to putout to the world and suddenly the
(01:45):
world embraced it.
So yes, I am excited about it, andI'm sure we'll come on to some of
the reasons why one needs toapproach it with caution.
But I think you need to put it inthe same category as the
internet, stroke, World Wide Weband the cloud and smartphones.
In terms of a once in agenerational, uh,
technology that's going to have hugerepercussions for the business
(02:06):
world.
Yeah. I mean, I was thinking aboutit this morning and, um, why
has ChatGPT become sowell known compared to the other
tools? And I'm just wondering, youknow, what you might think about
that, Sarah? What? Why is that onereally captured the market?
People's imagination.
Why do people think ChatGPTis the only product in some cases?
I mean, I think some of that is todo with the marketing and
(02:28):
the hype about it.
There's been some really clever, um,messaging that that's made
everybody realise that it is a toolthat they can, you know, literally
log on to themselves, create anaccount, you're up and running.
Um, so, you know, I think just thesnowball effect of the marketing
and then, you know, someone has ago, does something great, tells
their friend they all want to getinvolved. They want to see what they
(02:50):
can generate, not necessarilythings that are kind of cutting
edge, sort of maybe commercialoutputs, but actually quite a lot of
fun. But it gets peopleunderstanding the art of the
possible.
We've seen all the exciting thingsthat tools can do, like images
and poems. And I was watching atelevision show over the weekend,
and a couple were using ChatGPTor another tool other tools
(03:12):
are available, um, identifywhere they should move to.
Um, out of London.
I thought it was quite interesting,but Chris, why is
this so exciting from a businessperspective?
I think Sarah's partly touched onthat. It's ease of use.
You know, it's technology for themasses.
It's something that went viral.
That what? That's how it went from,you know, zero to hundreds of
millions of users globallywithin a matter of months.
(03:35):
Um, and I think why it's so excitingis because so much of professional
work these days involveswritten outputs or visual outputs.
And, um, you know,for any business that's
got a substantial amount of writtenor visual output, it's
already potentiallytransformational. I say potentially
because, you know, your people doneed to be trained up to a certain
(03:56):
degree in how to use it and some ofthe controls around it.
But certainly for the more routinenote taking or research
based, uh, scenarios,it's already a game changer.
We're seeing it very extensively inRSM for that.
Other industries, such asadvertising, the legal profession,
for instance.
It's already been a game changer.
We should put that on the table.
But as the technology advances, asit surely will, then a lot more use
(04:20):
cases will come into play.
Where are we on that that gamechanging journey?
You know, if we're going to usewords like hype cycle and use cases,
um, dare I say it like theidea that we use the technology to
just take notes and summarisedoesn't seem as exciting as some
of the things we've promised.
So where are we on that journey fromyour perspective?
Yeah, a couple of thoughts on that.
Then, I suppose, first of all, thethe most productivity enhancing
(04:43):
use cases aren't always the mostexciting.
Now there is a lot of routinestuff data entry
and manual copy creationand so forth that's done across
a lot of professions thatactually, if I
can do that, then you're freeing upa huge amount of, um,
intellectual time to devoteto, let's say, more value adding
(05:05):
things like winning new business or,um, coaching staff or whatever.
So even though there's someunsexy use cases that it started
with, actually those can bemassively productivity enhancing.
But also it's not like softwarethat you buy from a vendor that
goes through a gradual upgradein what it can do every few years.
(05:25):
The capability of generative AIis is increasing exponentially every
month with the release ofnew language models and and so
forth. So it's not a staticthing. Even since ChatGPT
was released, it's already far morecapable than it was then.
Yeah. Um, what do you think, Sarah?
Well, I think, you know, Chris isabsolutely right.
(05:47):
And I think, um, youneed to start somewhere that people
can get their heads round.
So the things we're seeing now arealready embedding this kind of
new way of working into thingsthat people do in their daily
lives.
Which helps them to get their headsround, actually, some of their
work is now being done by a machine.
How do I make that part of what I'mdoing and feel comfortable with it?
(06:10):
Because it's only going to evolve.
It's only going to become amore integral part of daily
lives, both, you know, professionaland personal.
But people do need to be educated.
And so I think what we're seeing nowis those foundational steps that
people will get used to using thiskind of technology, and then we can
build on that with the really moreadvanced use cases and kind of
(06:31):
evolve that as the technologyevolves.
But do you think people are boredyet? So just to articulate behind me
that I was in California and we talkabout it, some people seem fatigued
by the topic.
So I was just curious if you thinkthat's just my experience of the
conversation I had last month, or ifyou think actually it differs, like
where are people out?
I don't think people are bored.
(06:52):
I think, you know that there is alot of hype.
And yes, I've seen a few people havesaid, oh, you know, talking about AI
again, but actually they want tolearn. They want to understand what
it means for them and they want tounderstand how to get started.
So yes, again, that the publiclyavailable, the ChatGPT used
the things about the clause that areout there.
It's a great way to get started, butyou can see that organisations in
(07:15):
particular are starting to thinkthat how can I take that to the
next level? How can I actually makethat a bit more personal to me?
So I don't think they're bored, butI think there's a lot of curiosity,
a lot of questions.
Yeah, I think partly because it'sone of those technology
domains that spans both the personaland the business domain, doesn't it?
Most people or many peopleare using Intel on their
(07:38):
smartphones.
You know, they're using it forrecipes
or gym workout schedules orwhatever it happens to be.
And I think that builds bothknowledge and experience of how to
use it.
And, uh, no, noexcitement or enthusiasm is
the right phrase, but just itbecomes part of how you work.
It becomes part of what you do andhow you produce things.
(07:58):
Now, rather than having to startfrom scratch, you can use
generative AI.
And I think a lot of peopleare getting quite familiar with
that, actually. So, you know, forall those reasons, I think that this
is going to become second nature tomost professional workers within a
very short space of time.
It's probably a good time to talkabout what we're doing here at RSM
because we are using it.
(08:18):
So, Chris, I mean, are youhappy to share all.
Journey and our experiences and ourapproach of what we've done around
this area, and perhapstouched on some of the key things
we've learned.
Absolutely, Ben.
I mean, I think the first thing tosay is that the accounting
profession of which were part hasn'texactly been known in the past of
being at the leading edge oftechnology adoption.
But in this case, at RSM, wedecided that we had to be we
(08:41):
had to be exploring this, engagingour people with this, understanding
the risks.
And so back in March, April,I think it was Sarah, wasn't it?
Yeah.
We put together a a controlpilot.
So this is pretty soon afterChatGPT, you know, made such
a splash that weinvited volunteers across the firm.
We had, I think, about 100 peopleparticipate in that six week trial.
(09:06):
And we tried to strike a balancebetween bottling and distilling
and capturing how theyfound it, what worked, what didn't,
and just letting them get on withit. You know the word some
guardrails, let's say, aroundit. The main one being don't produce
client deliverables usingChatGPT.
So do research, do the backgroundprep work let's say.
(09:27):
Um, and what was really interestingactually, was the number of
surprising use cases that droppedout of that.
For instance, Excel macro creation.
Now that's one that when you thinkabout it, it's quite obvious because
it's quite a structured thing thatgenerative AI is really good at.
And an accounting firm, we've gotlots of Excel spreadsheets and so
forth. So we went through thatprocess for about six weeks
(09:48):
and um, used that then tobuild confidence to roll out an
enterprise wide secure version ofChatGPT, which of course is being
chat enterprise, which has come offthe back of Microsoft's investment
in open AI, which ensuresthat the prompts do go back into the
public domain.
And from then, we've rolled forwardinto, um, Microsoft
co-pilot and trialling that.
(10:08):
So we've taken what I describeas a a rapid semi-structured
approach, trying to get sort ofdemocratic, um,
experimentation going on, butwith some controls and guardrails.
That there's loads in there.
I mean, just to share, like myexpertise was in the pilot.
Um, and I remember, um, itwas the unexpected things that came
out. So a team member using it to,um, take their very technical
(10:33):
report, writing and then usethe tool to make it into something
which is closer to kind of moresimpler English.
And actually, um, you know, that menreflect and think is sometimes what
you don't know is the exciting thatcomes out of the these trials.
I mean, what what do you think,Sarah? What have you seen at RSM
that is useful to share?
Like maybe things that haven't quiteworked as well.
(10:54):
Yeah. I was going to start with,with one that I think is a great
use case for, for us here.
And, you know, to Chris's pointabout, you know, we produced a lot
of reports, a lot of documentationfor our clients, um,
not necessarily gettingChatGPT or being to enterprise
to create the report, butactually to validate
(11:15):
what we've written.
So to actually go through and ifwe've we've created an executive
summary at the beginning of ourreport, just ask.
The Gen AI to createan exact summary and just
cross-check it against what we'veput, make sure we're absolutely
on the mark with what we've put,because we know that a lot of our
clients will read the exact summaryand maybe not go that much further.
(11:36):
So that needs to be really, reallywatertight.
So that's a great use case.
Things that we haven't hadsuch great success with.
I wouldn't say it's not greatsuccess, but I think it's the trial
and error nature of Gen I.
And sometimes, you know, people areasking, they're putting in prompts
and they're not quite getting theresponse that they're looking for
(11:58):
and so they need to justtweak those prompts and craft
them in a different way.
And I suppose there's that elementof perseverance.
You know, we don't want people togive up because the first response
isn't quite the one that they werelooking for.
So there's a bit of education thatwe need to do there to make sure
people are getting the mostout of the tool.
What does that look like?
The people which are important partof this was the prompting and how
(12:21):
they feel. So how do we preparepeople for Gen AI?
We've got a number of, um, tracksthat we're following in this area,
um, being led by our, um,learning and development team.
So we've got a sort of digitaltraining manager, and she's looking
at a combination of our owncommunication and making
sure that we've got, um, regularcomms going out so that we kind
(12:45):
of sweep people up and keep sweepingmore and more people up in the wave,
because not everybody isconsuming information that comes
out to them when it first comes out.
So kind of a continual stream ofcommunication, working
on some basic, um,training material to help
people with actuallythe most direct prompting
(13:06):
that they can use.
Um, because we know that Gen AIis really precise, it doesn't
understand nuances.
You need to tell it exactly what youwant to get, the answer that you're
looking for, looking at training tohelp people really refine their use
of the English language, actually.
Which I think it's quite fascinating.
Just one other thing I'll throw in.
That's been something that I thinkis a bit of a limitation so far.
(13:28):
You know, we focus in theconversation so far largely been on
natural language outputs.
But of course gen AI is terrificin some cases and projecting images
and so forth.
But it's not quite there yet withregard to the quality of
the imagery that you can includein proposals or
reports, and partlybecause all large firms and
(13:51):
RSM is no exception, protect ourbrand.
So brand compliance of the imagesis something that, um,
Microsoft and the other peopleworking with Gen AI solutions are
still going to get their head roundand and still try to I think they're
going to realise that for us to haveconfidence in AI generated
visual outputs, it's going to bebrand compliant and there isn't
currently a way to do that.
(14:12):
I don't think that reallystreamline the process of, let's
say, proposal creation.
And also I think there are ethicalquestions around that too.
We have tried it in some ofour informal staff meetings,
for instance, you know, go off andprepare a presentation using
generative AI on whatever topic,and they come back and all of the
people that have been depicted aswhite, you know, and so there's the
(14:34):
ethical considerationsaround the adoption of the visual
outputs of AI that businessesneed to be more comfortable about
than they are at the moment.
It's a funny point you make aboutthe tools because to share an
example, I asked it to make apresentation on artificial
intelligence. Yeah.
And um, it produced the textand I said, oh, can you make it more
succinct? And it seemed to be ableto handle that.
(14:56):
And then I asked, how can you makeit more visual?
And all it did was place an enormousphoto of a semiconductor chip across
things. And I was like, well, that'snot really what I need.
And I see the potential of thetechnology.
Yeah. And I guess my question iswhen you experience something like
that in the workforce, what dobusinesses what do we need to do to
keep people engaged with it so theydon't disengage with it?
(15:19):
Probably quite an important point.
Um, that's an interesting question.
Maybe it comes down to nothaving a continual stream of
pressure to try it, try it, try it,because the technology will
evolve in steps.
So a specific example ofthis, some of them are
good for language and some otherones are good for images
(15:40):
until you get tools that canintegrate both.
I think we're going to struggle tocreate the types of
artefacts that, um,businesses need, because
very few documents these days arejust text.
They've often got imagesincorporated, or maybe a video
content.
And I think the Holy Grail is agenerative AI that you can prompt
(16:00):
with. Create me a documentthat does whatever
you need it to do, but the outputsare, um, multimedia
that we're not seeing yet, andwe're seeing that in some of the
trials with Microsoft Co-Pilot, forinstance, which is terrific in many
use cases, particularly virtualmeeting summarisation potential game
changer there.
But when it comes to on brandPowerPoint.
(16:23):
Presentations.
It's not there yet because of thelack of interaction with corporate
brand templates, but also becauseI think its strength is really in
the natural language outputs,rather than the visual outputs
rather than both.
What do you think, Sarah people, Ialways talk about people and AI and
tools, because I do feellike an important part is
interaction of how we work withthese tools.
(16:45):
But I guess same question to you.
If it's not quite working, the magicisn't there.
How do we keep people engaged withit so they continue to experiment,
iterate, and find new ideas?
I think, um, we do needto keep encouraging people to
keep experimenting, but Ithink we need to start to be, um,
maybe a little bit more focussed.
So, you know, we're using publiclyavailable tools at the moment, which
(17:08):
is great.
Um, but we actually don't reallyknow exactly what the underlying
data is in those tools.
We know it's come from the WorldWide Web, but how
much of that content's beenmoderated? What's been selected to
be part of the model?
We don't know.
And so, um, therefore we can'tbe sure of what the outputs are
going to be.
(17:28):
So I think what we need to be doingis encouraging people to think about
slightly narrower applicationsof generative AI that are really
specific to what they do on a dailybasis, and then identifying
the data that would helpthem to achieve what they're trying
to achieve. So actually just reallybe a bit more focussed and a bit
more specific.
(17:48):
And I think people will feel moreengaged because it's more relevant
to them.
Mhm. I think that's a really goodpoint.
And um, I think thatall businesses are going to find
that their staff will graduallybecome accustomed to using the
everyday generative AI, the publiclyavailable ones, ChatGPT, Dall-E and
all the rest of it.
It's how you then use the languagemodels that are out there to create
(18:11):
specific use cases in your business,using your corporate data.
To Sarah's point about the trainingdata, which is essential.
You've got more control over that,but also that's how you can
differentiate because it's yourdata.
Your data is such a huge asset,and that's how you can leverage the
value in that data is byconnecting up to those language
(18:32):
models, um, that ahandful of organisations around
the world found the resource todevelop, and there is only still a
handful, let's face it.
But any business can connect tothose, and they need the
software developer and data scienceskills to do so.
Apply those language models to theircorporate data sets and that's where
it gets really exciting.
Should we talk about data?
(18:53):
I mean, data is not a new word.
It's been around a long time.
But so like what's the relationshipbetween data and AI?
And then why is it so important tounderstand that relationship.
So I think to the point we were justmaking all of these language
models and any traditional AImodels are based on data.
And so, um, the outputsof the AI that you're going to get
(19:17):
are only as good as the datathat's been input to the models.
Um, and I'm sure that, you know,the majority of people are aware of
that.
And in fact, you know, data has beenessential for driving many
things in organisations.
So analytics, automation, nowAI. They all require a solid
data foundation.
(19:37):
I think the difference with AI isbecause it's so powerful,
it can also be very powerfullywrong if the data is wrong.
So the need for really good qualitydata is kind of
exacerbated, if you like, by the AI.
Has the AI said the rise of genAI? Has that changed the way we look
at data, or whether the businessshould look at data?
(19:57):
Or is it the same challengebut just through a different lens?
What's your perspective on that?
I personally think it's the samechallenge through a different lens,
and I think, you know, organisationsshould, if they haven't already,
be thinking about or have agood solid data strategy
so that they understand the datathey have.
They understand where it resides,how it flows through their
(20:20):
organisation, who owns it.
They have, um, arealistic view of the quality
of their data, and they put theright processes and controls
in place to actually govern theirdata and
think about how they want to usedata.
You know, ten years ago it was allabout big data.
Let's just get all our data togetherand let's, you know, there's loads
(20:43):
of value in data, which of coursethere is.
But it wasn't very specificin terms of what do you actually
want to do with your data.
So that a data strategywith some really concrete,
um, goals, what am I going todo? What am I trying to achieve with
my data? Whether it's an analyticsuse case or an AI use case
is really, really fundamentalbecause you can go off
(21:06):
in lots of directions andmaybe not quite achieve
the value you or you could achieve,um, without kind of the strategy
and the vision in place.
It just makes me wonder, Chris,like, how do you stay focussed?
How do you avoid that?
What maybe look like like ascattergun approach of data?
What's your kind of perspective onhow would a a board or
business stay focussedwhen tackling data?
(21:29):
I think that is the risk benefit ofa scattergun approach to this.
And I think, um, that'sexacerbated by the fact that when we
talk about data, we're not justtalking about structured data here.
Talk about unstructured data,content reports, videos,
um, libraries of images, allof those are part of an
organisation's, uh, data,and they've got value to them.
(21:50):
And generative AI can learn fromand be trained on them.
You know, if you're a creativeagency, then it isn't so much
about the wordsthat you've produced in the past.
That's your IP, it's the visuals,it's the content.
It's it's the videos.
And generative AI can increasinglylearn from that as long as it's got
the right metadata attached to it, Isuppose.
(22:11):
Um, but you're absolutely right.
You know, you've got to avoid ascattergun approach, which means
fundamentally, I think,understanding how the technology
works.
You know, you've got to have peoplein your organisation that really
understand how to get the bestout of a generative AI model
and, uh, how to feed data intoit, uh, how to experiment with
the outputs, how to avoidhallucinations, how to manage the
(22:33):
security and the ethics and theprivacy considerations around it
all.
Um, so it does, again, as most,uh, technology adoption plans
do come down to having the rightskills in place to get the right
focus.
It just feels like a lot to thinkabout.
To me, the more I think about it andit's there. But, um,
taking that step back, whereare the key areas leadership teams
(22:55):
need to start at this point?
Like what they need to think aboutto get.
You've touched on actually with theapproach, right. But where's
the absolute starting point witha blank piece of paper?
I think the mistake would be formost organisations to say, right,
we need an AI strategy.
We're going to create the perfect AIstrategy. Now, actually,
you do need to experiment.
You need to do the kind of thingswe've been talking about controlled
(23:17):
trials to better understand theskills, the safeguards and the
use cases for your business.
That could then form the kernel ofan AI strategy for the future,
because the AI that we use now willbe the worst it will ever use,
because it will very quickly getbetter and better and better.
So an AI strategy for today.
If you were to waste your timetrying to create the perfect
(23:38):
document that is your AI strategywill be out of date in a couple of
years, if not sooner.
So thinking about how AIand generative AI particularly
is going to affect your businessmodel is a bit of a
fool's game unless you're in thegenerative AI business, or maybe
the software business for the restof us.
Start by trying it out.
(23:58):
Get your staff comfortable withprompting, and start to bring
in the kind of data science.
Skills that you'll need to thendifferentiate with AI.
That's what I think mostbusinesses should be doing.
That's what I'm advising our clientsto do. Unless you've already got
very, very deep data science skills.
Can we go back on control trial?
And, um, what I'm like for a pilot,but I'd be interested to hear
(24:20):
thoughts from both of you. What doesa good control trial look like?
So I think there's a bit ofa balance here because as Chris
said, you know, you need people toexperiment. You don't want to be too
prescriptive about what you wantthem to do, but you do
need to be able tomeasure what they're doing.
You do need some feedback.
(24:40):
You've got to create forums wherebypeople can share their experiences,
where they can provide feedback,where they can evaluate the
functionality and share what'sgood and what's not good, and
work together and build on that.
If you just set up a trial and say,you know, see you in a year,
you're not going to get any valuefrom it.
And you do need someone tocoordinate that, don't you?
(25:01):
We've got a full time digitalcommunity coordinator, for instance.
Even he is snowed with trying tocoordinate all these different
professional workers and, you know,get what works and what doesn't and
get them to, you know, create theirtips and tricks and so forth.
I think also, you've got to involverisk in compliance from the start in
a trial, don't you.
Mhm. Um, you know you've got to havesome guardrails
(25:22):
is the kind of word that everyone'stalking about which simply mean some
do's and don'ts.
Don't use personal data.
Don't use it to create client orcustomer facing outputs for now
until we better understand some ofthe risks around it.
But do use it for anythinginternally that is associated
with research or meeting prep,or things like coding and,
(25:43):
uh, Excel macros and the like, whichwe've mentioned, which are much less
risky because codeis not something that's covered by
intellectual property.
As such, it either works or itdoesn't.
So, um, I think havingthose guardrails that people need to
understand is a key part of a trial.
There's a cultural point to, isn'tthere, like the nature
(26:03):
of a lot of people in the accountingprofession is they see this this
tool and they want to break it.
They'll ask it a question like, tellme how this tax legislation works.
Tell me how this IFRS applied andwhen it doesn't work, sort of sit
back and say, oh, we're safe.
It doesn't work.
Yeah. But actually part ofthe culture has to be to take
all the narrative and all theconversation around that testing and
(26:24):
make it constructive, whichtakes time.
People need to feel safe.
There needs to be some joy in it,doesn't there?
I don't know, that's how I feel.
But be curious how you feel Sarahabout the culture.
How do you get people playing withit that is constructive?
Yeah, and it sort of plays backto the point about the digital
community and havingpeople who feel part of
(26:47):
an environment where they feel safeto do that experimentation,
where they've got support if theyneed it, because, um,
sometimes just a little bitof a pointer in the right direction
can make a really big difference forwhatever it is somebody trying to
do. So, you know, havingthose collaboration helps
that community feel,um, but I think, you know,
(27:09):
on a broader level around theculture, um, all
of this does go hand in hand withbasic levels of data literacy,
which are part of agood data strategy, is, you know,
it can't just be about thetechnology, the tools, the data.
It needs to be about how peopleinteract with those things, how
(27:29):
they feel comfortable and getconfidence in the inputs
and the outputs, howthey feel empowered to challenge,
to to question when they thinksomething might not be right.
And all of that comes back to basicdata literacy and just asking
questions of where did that datacome from? How did it come up with
that answer?
Am I sure it's the right answer?
(27:50):
How do I cross-check?
And that's just a big educationpiece and it doesn't happen
overnight. You have to kind of keepgoing with that.
I think also peopleare used to the world of software
either working or not working, butwith generative AI, the shades
of grey, it worksvery well most of the time, but
sometimes it hallucinates andoccasionally it just gets things a
(28:11):
bit muddled up or whatever.
But human beings are fallible too,and so I think that it's
got the reason it reached aninflection point.
I think after years of developmentof, um, language models
and so forth, was because itsuddenly became
as good in many scenarios as a humanbeing, because we're all fallible,
we can put errors into ourinto our written outputs and so
(28:33):
forth.
And so it's it's not ablack or white, it works or it
doesn't.
Largely it's good enough for many,many use cases.
So we shouldn't be waiting forperfection.
We should be using it now.
But with the right review around it,you know, it's an assistant isn't
it? It's assistive technology.
And as long as you review what itproduces, we should be using it
(28:55):
now.
So a year ago, broadly,no one had heard of Gen AI.
I know people had, but generalrecognition was very, very low.
I think now it's fair to say mostpeople have heard of it and played
with it. The two questionswhere will we be in 5 or 10 years
time?
Who wants to tackle five years andwho wants to tackle ten years?
Uh, that's sohard, isn't it? Because it's
(29:18):
evolving at such a pace thatI think it's pretty much impossible
to predict specifically what it'llbe capable of in our own profession.
There's a lot of assumptions aroundthe types of compliance
work, the types of advisory workthat a human being is
essential for.
Now, I still think that human beingswill be essential for most of that
(29:38):
work in the future, but in muchsmaller ways.
Um, and so I thinkthat the ability
of generative AI to incorporatea lot more data from the real world
that go into all of ourthought processes when we as
human beings are producing, um,written or visual outputs
(30:00):
can be incorporated into the modelsin the future, and it will be able
to do a lot of things that humanbeings currently assume that they
are. Um, you know, thatthe masters of for now.
So, I mean, if I'm going to make aprediction, is that a lot of the
current assumptions around what ahuman being is needed for will be
gone.
But in terms of specifics,if only I knew Ben!
(30:23):
You mentioned it.
A lot will change.
But is it too early to reflectdeeply on the impact of this
technology because it's so muchunknown?
I don't think it is.
I don't think it is. I mean, we youknow, there's a lot of examples of
generative AI in our daily livesthat a lot of people don't even
realise is, is AI enabled,you know, the Siri and the, um,
Amazon Alexa and all those kind ofthings, which, again, are getting
(30:44):
better and better all the time.
Um, so I think it's aboutgenerative AI being much
more woven into the fabric ofour daily lives.
You know, in cars, in,uh, bus stations and train
stations, in airports as well asin the working environments.
I think it's going to be a case ofit will augment most of what we
(31:06):
do in our daily lives, personallyand professionally.
That's what we need tostart thinking about now.
And obviously, the Bletchley Parksummit recently was a part of that
to begin to think about the privacy,um, and regulatory aspects
of it.
It's definitely not too soon to bethinking about that.
We've got to be thinking about an AIaugmented, you know, future
(31:27):
of humanity candidly.
But to bring it back to the businessworld.
All businesses are going to be AIenhanced, or most businesses already
are, even if they don't realise it,but becoming more explicitly so in
the next few years.
Yeah, and I think to build onthat, you know, the first
phase, I'm not going to put years onit, but just the first phase is
definitely about embeddinga different way of working and a
(31:49):
different way of thinking, uh, youknow, across personal lives
and professional lives.
Um, I think what we'll see,let's say, within the 5 to 10 year
window is a lot more proprietary,um, AI, generative
AI and AI models whereorganisations invest
in building something that is very,very personal to them and their
(32:11):
workforce.
But that isn't that can't happenquickly. It's not going to happen
overnight. It will take a lot ofinvestment. It's quite complicated.
So I would counsel that,organisations don't
rush into that.
Start thinking about what thoseuses, use cases might be,
what would be the underlying datathat's needed.
Get your workforce ready andused to working with this kind
(32:33):
of technology, and then investin the right use case.
And it might just be one.
Make sure it's the one that's goingto bring the most value.
Thank you so much.
Um, Chris and Sarah, um,if you would like to learn more
about Genrative AI and someinsights, please take a look at our
website and we'll put a link to thatin the show notes.
And please look out for furtherepisodes of The Loop, where we'll
(32:54):
continue to investigate the impactof this technology on businesses.
And thank you very much forlistening.
Hello and welcome to The Loopwith ChatGPT recently celebrating
its first birthday.
We're going to be discussinggenerative artificial intelligence.
So joining me,our RSM chief digital officer, Chris
Knowles anddata analytics and insights partner,
Sarah Belsham.
So, Sarah, Chris, welcome.
Thank you. Great to be here.
(00:20):
Thank you.
So nobody had heard of generativeartificial intelligence until
ChatGPT.
So Sarah, whyhas Gen I taken on so much momentum
compared to all the other AI toolsthat have been out there for a
number of years?
Yeah, it's an interesting point Ben,because obviously AI isn't
a new technology, not by any means.
It's been around for, you know, 50plus years.
(00:42):
Um, but I think the traditional AIthat, um, people may
be aware of requiresquite a lot of extensive compute
and processing power, largevolumes of data, um, and
actually quite a lot of veryspecialist skills around sort
of statistical analysis of data.
So it's not for the masses byany stretch.
(01:03):
Whereas ChatGPT launchedand offered a very different flavour
of AI, something much moreaccessible to people and something
more akin to what people are usedto in their day to day lives, more
intuitive and not necessarilyrequiring the level of technical
skills that traditional AI requires.
Chris, do you do you get excitedabout Gen AI year on from ChatGPT?
(01:25):
I am excited, and I think, as Sarahsaid, it's, uh, you know, it's
building on a technology that's beenaround for a while, but it's it
reached a tipping point, didn't it,where the usability, the quality of
the natural language output becamesuch that suddenly it was
something that OpenAI, inconjunction with Microsoft, as we
know, was confident enough to putout to the world and suddenly the
(01:45):
world embraced it.
So yes, I am excited about it, andI'm sure we'll come on to some of
the reasons why one needs toapproach it with caution.
But I think you need to put it inthe same category as the
internet, stroke, World Wide Weband the cloud and smartphones.
In terms of a once in agenerational, uh,
technology that's going to have hugerepercussions for the business
(02:06):
world.
Yeah. I mean, I was thinking aboutit this morning and, um, why
has ChatGPT become sowell known compared to the other
tools? And I'm just wondering, youknow, what you might think about
that, Sarah? What? Why is that onereally captured the market?
People's imagination.
Why do people think ChatGPTis the only product in some cases?
I mean, I think some of that is todo with the marketing and
(02:28):
the hype about it.
There's been some really clever, um,messaging that that's made
everybody realise that it is a toolthat they can, you know, literally
log on to themselves, create anaccount, you're up and running.
Um, so, you know, I think just thesnowball effect of the marketing
and then, you know, someone has ago, does something great, tells
their friend they all want to getinvolved. They want to see what they
(02:50):
can generate, not necessarilythings that are kind of cutting
edge, sort of maybe commercialoutputs, but actually quite a lot of
fun. But it gets peopleunderstanding the art of the
possible.
We've seen all the exciting thingsthat tools can do, like images
and poems. And I was watching atelevision show over the weekend,
and a couple were using ChatGPTor another tool other tools
(03:12):
are available, um, identifywhere they should move to.
Um, out of London.
I thought it was quite interesting,but Chris, why is
this so exciting from a businessperspective?
I think Sarah's partly touched onthat. It's ease of use.
You know, it's technology for themasses.
It's something that went viral.
That what? That's how it went from,you know, zero to hundreds of
millions of users globallywithin a matter of months.
(03:35):
Um, and I think why it's so excitingis because so much of professional
work these days involveswritten outputs or visual outputs.
And, um, you know,for any business that's
got a substantial amount of writtenor visual output, it's
already potentiallytransformational. I say potentially
because, you know, your people doneed to be trained up to a certain
(03:56):
degree in how to use it and some ofthe controls around it.
But certainly for the more routinenote taking or research
based, uh, scenarios,it's already a game changer.
We're seeing it very extensively inRSM for that.
Other industries, such asadvertising, the legal profession,
for instance.
It's already been a game changer.
We should put that on the table.
But as the technology advances, asit surely will, then a lot more use
(04:20):
cases will come into play.
Where are we on that that gamechanging journey?
You know, if we're going to usewords like hype cycle and use cases,
um, dare I say it like theidea that we use the technology to
just take notes and summarisedoesn't seem as exciting as some
of the things we've promised.
So where are we on that journey fromyour perspective?
Yeah, a couple of thoughts on that.
Then, I suppose, first of all, thethe most productivity enhancing
(04:43):
use cases aren't always the mostexciting.
Now there is a lot of routinestuff data entry
and manual copy creationand so forth that's done across
a lot of professions thatactually, if I
can do that, then you're freeing upa huge amount of, um,
intellectual time to devoteto, let's say, more value adding
(05:05):
things like winning new business or,um, coaching staff or whatever.
So even though there's someunsexy use cases that it started
with, actually those can bemassively productivity enhancing.
But also it's not like softwarethat you buy from a vendor that
goes through a gradual upgradein what it can do every few years.
(05:25):
The capability of generative AIis is increasing exponentially every
month with the release ofnew language models and and so
forth. So it's not a staticthing. Even since ChatGPT
was released, it's already far morecapable than it was then.
Yeah. Um, what do you think, Sarah?
Well, I think, you know, Chris isabsolutely right.
(05:47):
And I think, um, youneed to start somewhere that people
can get their heads round.
So the things we're seeing now arealready embedding this kind of
new way of working into thingsthat people do in their daily
lives.
Which helps them to get their headsround, actually, some of their
work is now being done by a machine.
How do I make that part of what I'mdoing and feel comfortable with it?
(06:10):
Because it's only going to evolve.
It's only going to become amore integral part of daily
lives, both, you know, professionaland personal.
But people do need to be educated.
And so I think what we're seeing nowis those foundational steps that
people will get used to using thiskind of technology, and then we can
build on that with the really moreadvanced use cases and kind of
(06:31):
evolve that as the technologyevolves.
But do you think people are boredyet? So just to articulate behind me
that I was in California and we talkabout it, some people seem fatigued
by the topic.
So I was just curious if you thinkthat's just my experience of the
conversation I had last month, or ifyou think actually it differs, like
where are people out?
I don't think people are bored.
(06:52):
I think, you know that there is alot of hype.
And yes, I've seen a few people havesaid, oh, you know, talking about AI
again, but actually they want tolearn. They want to understand what
it means for them and they want tounderstand how to get started.
So yes, again, that the publiclyavailable, the ChatGPT used
the things about the clause that areout there.
It's a great way to get started, butyou can see that organisations in
(07:15):
particular are starting to thinkthat how can I take that to the
next level? How can I actually makethat a bit more personal to me?
So I don't think they're bored, butI think there's a lot of curiosity,
a lot of questions.
Yeah, I think partly because it'sone of those technology
domains that spans both the personaland the business domain, doesn't it?
Most people or many peopleare using Intel on their
(07:38):
smartphones.
You know, they're using it forrecipes
or gym workout schedules orwhatever it happens to be.
And I think that builds bothknowledge and experience of how to
use it.
And, uh, no, noexcitement or enthusiasm is
the right phrase, but just itbecomes part of how you work.
It becomes part of what you do andhow you produce things.
(07:58):
Now, rather than having to startfrom scratch, you can use
generative AI.
And I think a lot of peopleare getting quite familiar with
that, actually. So, you know, forall those reasons, I think that this
is going to become second nature tomost professional workers within a
very short space of time.
It's probably a good time to talkabout what we're doing here at RSM
because we are using it.
(08:18):
So, Chris, I mean, are youhappy to share all.
Journey and our experiences and ourapproach of what we've done around
this area, and perhapstouched on some of the key things
we've learned.
Absolutely, Ben.
I mean, I think the first thing tosay is that the accounting
profession of which were part hasn'texactly been known in the past of
being at the leading edge oftechnology adoption.
But in this case, at RSM, wedecided that we had to be we
(08:41):
had to be exploring this, engagingour people with this, understanding
the risks.
And so back in March, April,I think it was Sarah, wasn't it?
Yeah.
We put together a a controlpilot.
So this is pretty soon afterChatGPT, you know, made such
a splash that weinvited volunteers across the firm.
We had, I think, about 100 peopleparticipate in that six week trial.
(09:06):
And we tried to strike a balancebetween bottling and distilling
and capturing how theyfound it, what worked, what didn't,
and just letting them get on withit. You know the word some
guardrails, let's say, aroundit. The main one being don't produce
client deliverables usingChatGPT.
So do research, do the backgroundprep work let's say.
(09:27):
Um, and what was really interestingactually, was the number of
surprising use cases that droppedout of that.
For instance, Excel macro creation.
Now that's one that when you thinkabout it, it's quite obvious because
it's quite a structured thing thatgenerative AI is really good at.
And an accounting firm, we've gotlots of Excel spreadsheets and so
forth. So we went through thatprocess for about six weeks
(09:48):
and um, used that then tobuild confidence to roll out an
enterprise wide secure version ofChatGPT, which of course is being
chat enterprise, which has come offthe back of Microsoft's investment
in open AI, which ensuresthat the prompts do go back into the
public domain.
And from then, we've rolled forwardinto, um, Microsoft
co-pilot and trialling that.
(10:08):
So we've taken what I describeas a a rapid semi-structured
approach, trying to get sort ofdemocratic, um,
experimentation going on, butwith some controls and guardrails.
That there's loads in there.
I mean, just to share, like myexpertise was in the pilot.
Um, and I remember, um, itwas the unexpected things that came
out. So a team member using it to,um, take their very technical
(10:33):
report, writing and then usethe tool to make it into something
which is closer to kind of moresimpler English.
And actually, um, you know, that menreflect and think is sometimes what
you don't know is the exciting thatcomes out of the these trials.
I mean, what what do you think,Sarah? What have you seen at RSM
that is useful to share?
Like maybe things that haven't quiteworked as well.
(10:54):
Yeah. I was going to start with,with one that I think is a great
use case for, for us here.
And, you know, to Chris's pointabout, you know, we produced a lot
of reports, a lot of documentationfor our clients, um,
not necessarily gettingChatGPT or being to enterprise
to create the report, butactually to validate
(11:15):
what we've written.
So to actually go through and ifwe've we've created an executive
summary at the beginning of ourreport, just ask.
The Gen AI to createan exact summary and just
cross-check it against what we'veput, make sure we're absolutely
on the mark with what we've put,because we know that a lot of our
clients will read the exact summaryand maybe not go that much further.
(11:36):
So that needs to be really, reallywatertight.
So that's a great use case.
Things that we haven't hadsuch great success with.
I wouldn't say it's not greatsuccess, but I think it's the trial
and error nature of Gen I.
And sometimes, you know, people areasking, they're putting in prompts
and they're not quite getting theresponse that they're looking for
(11:58):
and so they need to justtweak those prompts and craft
them in a different way.
And I suppose there's that elementof perseverance.
You know, we don't want people togive up because the first response
isn't quite the one that they werelooking for.
So there's a bit of education thatwe need to do there to make sure
people are getting the mostout of the tool.
What does that look like?
The people which are important partof this was the prompting and how
(12:21):
they feel. So how do we preparepeople for Gen AI?
We've got a number of, um, tracksthat we're following in this area,
um, being led by our, um,learning and development team.
So we've got a sort of digitaltraining manager, and she's looking
at a combination of our owncommunication and making
sure that we've got, um, regularcomms going out so that we kind
(12:45):
of sweep people up and keep sweepingmore and more people up in the wave,
because not everybody isconsuming information that comes
out to them when it first comes out.
So kind of a continual stream ofcommunication, working
on some basic, um,training material to help
people with actuallythe most direct prompting
(13:06):
that they can use.
Um, because we know that Gen AIis really precise, it doesn't
understand nuances.
You need to tell it exactly what youwant to get, the answer that you're
looking for, looking at training tohelp people really refine their use
of the English language, actually.
Which I think it's quite fascinating.
Just one other thing I'll throw in.
That's been something that I thinkis a bit of a limitation so far.
(13:28):
You know, we focus in theconversation so far largely been on
natural language outputs.
But of course gen AI is terrificin some cases and projecting images
and so forth.
But it's not quite there yet withregard to the quality of
the imagery that you can includein proposals or
reports, and partlybecause all large firms and
(13:51):
RSM is no exception, protect ourbrand.
So brand compliance of the imagesis something that, um,
Microsoft and the other peopleworking with Gen AI solutions are
still going to get their head roundand and still try to I think they're
going to realise that for us to haveconfidence in AI generated
visual outputs, it's going to bebrand compliant and there isn't
currently a way to do that.
(14:12):
I don't think that reallystreamline the process of, let's
say, proposal creation.
And also I think there are ethicalquestions around that too.
We have tried it in some ofour informal staff meetings,
for instance, you know, go off andprepare a presentation using
generative AI on whatever topic,and they come back and all of the
people that have been depicted aswhite, you know, and so there's the
(14:34):
ethical considerationsaround the adoption of the visual
outputs of AI that businessesneed to be more comfortable about
than they are at the moment.
It's a funny point you make aboutthe tools because to share an
example, I asked it to make apresentation on artificial
intelligence. Yeah.
And um, it produced the textand I said, oh, can you make it more
succinct? And it seemed to be ableto handle that.
(14:56):
And then I asked, how can you makeit more visual?
And all it did was place an enormousphoto of a semiconductor chip across
things. And I was like, well, that'snot really what I need.
And I see the potential of thetechnology.
Yeah. And I guess my question iswhen you experience something like
that in the workforce, what dobusinesses what do we need to do to
keep people engaged with it so theydon't disengage with it?
(15:19):
Probably quite an important point.
Um, that's an interesting question.
Maybe it comes down to nothaving a continual stream of
pressure to try it, try it, try it,because the technology will
evolve in steps.
So a specific example ofthis, some of them are
good for language and some otherones are good for images
(15:40):
until you get tools that canintegrate both.
I think we're going to struggle tocreate the types of
artefacts that, um,businesses need, because
very few documents these days arejust text.
They've often got imagesincorporated, or maybe a video
content.
And I think the Holy Grail is agenerative AI that you can prompt
(16:00):
with. Create me a documentthat does whatever
you need it to do, but the outputsare, um, multimedia
that we're not seeing yet, andwe're seeing that in some of the
trials with Microsoft Co-Pilot, forinstance, which is terrific in many
use cases, particularly virtualmeeting summarisation potential game
changer there.
But when it comes to on brandPowerPoint.
(16:23):
Presentations.
It's not there yet because of thelack of interaction with corporate
brand templates, but also becauseI think its strength is really in
the natural language outputs,rather than the visual outputs
rather than both.
What do you think, Sarah people, Ialways talk about people and AI and
tools, because I do feellike an important part is
interaction of how we work withthese tools.
(16:45):
But I guess same question to you.
If it's not quite working, the magicisn't there.
How do we keep people engaged withit so they continue to experiment,
iterate, and find new ideas?
I think, um, we do needto keep encouraging people to
keep experimenting, but Ithink we need to start to be, um,
maybe a little bit more focussed.
So, you know, we're using publiclyavailable tools at the moment, which
(17:08):
is great.
Um, but we actually don't reallyknow exactly what the underlying
data is in those tools.
We know it's come from the WorldWide Web, but how
much of that content's beenmoderated? What's been selected to
be part of the model?
We don't know.
And so, um, therefore we can'tbe sure of what the outputs are
going to be.
(17:28):
So I think what we need to be doingis encouraging people to think about
slightly narrower applicationsof generative AI that are really
specific to what they do on a dailybasis, and then identifying
the data that would helpthem to achieve what they're trying
to achieve. So actually just reallybe a bit more focussed and a bit
more specific.
(17:48):
And I think people will feel moreengaged because it's more relevant
to them.
Mhm. I think that's a really goodpoint.
And um, I think thatall businesses are going to find
that their staff will graduallybecome accustomed to using the
everyday generative AI, the publiclyavailable ones, ChatGPT, Dall-E and
all the rest of it.
It's how you then use the languagemodels that are out there to create
(18:11):
specific use cases in your business,using your corporate data.
To Sarah's point about the trainingdata, which is essential.
You've got more control over that,but also that's how you can
differentiate because it's yourdata.
Your data is such a huge asset,and that's how you can leverage the
value in that data is byconnecting up to those language
(18:32):
models, um, that ahandful of organisations around
the world found the resource todevelop, and there is only still a
handful, let's face it.
But any business can connect tothose, and they need the
software developer and data scienceskills to do so.
Apply those language models to theircorporate data sets and that's where
it gets really exciting.
Should we talk about data?
(18:53):
I mean, data is not a new word.
It's been around a long time.
But so like what's the relationshipbetween data and AI?
And then why is it so important tounderstand that relationship.
So I think to the point we were justmaking all of these language
models and any traditional AImodels are based on data.
And so, um, the outputsof the AI that you're going to get
(19:17):
are only as good as the datathat's been input to the models.
Um, and I'm sure that, you know,the majority of people are aware of
that.
And in fact, you know, data has beenessential for driving many
things in organisations.
So analytics, automation, nowAI. They all require a solid
data foundation.
(19:37):
I think the difference with AI isbecause it's so powerful,
it can also be very powerfullywrong if the data is wrong.
So the need for really good qualitydata is kind of
exacerbated, if you like, by the AI.
Has the AI said the rise of genAI? Has that changed the way we look
at data, or whether the businessshould look at data?
(19:57):
Or is it the same challengebut just through a different lens?
What's your perspective on that?
I personally think it's the samechallenge through a different lens,
and I think, you know, organisationsshould, if they haven't already,
be thinking about or have agood solid data strategy
so that they understand the datathey have.
They understand where it resides,how it flows through their
(20:20):
organisation, who owns it.
They have, um, arealistic view of the quality
of their data, and they put theright processes and controls
in place to actually govern theirdata and
think about how they want to usedata.
You know, ten years ago it was allabout big data.
Let's just get all our data togetherand let's, you know, there's loads
(20:43):
of value in data, which of coursethere is.
But it wasn't very specificin terms of what do you actually
want to do with your data.
So that a data strategywith some really concrete,
um, goals, what am I going todo? What am I trying to achieve with
my data? Whether it's an analyticsuse case or an AI use case
is really, really fundamentalbecause you can go off
(21:06):
in lots of directions andmaybe not quite achieve
the value you or you could achieve,um, without kind of the strategy
and the vision in place.
It just makes me wonder, Chris,like, how do you stay focussed?
How do you avoid that?
What maybe look like like ascattergun approach of data?
What's your kind of perspective onhow would a a board or
business stay focussedwhen tackling data?
(21:29):
I think that is the risk benefit ofa scattergun approach to this.
And I think, um, that'sexacerbated by the fact that when we
talk about data, we're not justtalking about structured data here.
Talk about unstructured data,content reports, videos,
um, libraries of images, allof those are part of an
organisation's, uh, data,and they've got value to them.
(21:50):
And generative AI can learn fromand be trained on them.
You know, if you're a creativeagency, then it isn't so much
about the wordsthat you've produced in the past.
That's your IP, it's the visuals,it's the content.
It's it's the videos.
And generative AI can increasinglylearn from that as long as it's got
the right metadata attached to it, Isuppose.
(22:11):
Um, but you're absolutely right.
You know, you've got to avoid ascattergun approach, which means
fundamentally, I think,understanding how the technology
works.
You know, you've got to have peoplein your organisation that really
understand how to get the bestout of a generative AI model
and, uh, how to feed data intoit, uh, how to experiment with
the outputs, how to avoidhallucinations, how to manage the
(22:33):
security and the ethics and theprivacy considerations around it
all.
Um, so it does, again, as most,uh, technology adoption plans
do come down to having the rightskills in place to get the right
focus.
It just feels like a lot to thinkabout.
To me, the more I think about it andit's there. But, um,
taking that step back, whereare the key areas leadership teams
(22:55):
need to start at this point?
Like what they need to think aboutto get.
You've touched on actually with theapproach, right. But where's
the absolute starting point witha blank piece of paper?
I think the mistake would be formost organisations to say, right,
we need an AI strategy.
We're going to create the perfect AIstrategy. Now, actually,
you do need to experiment.
You need to do the kind of thingswe've been talking about controlled
(23:17):
trials to better understand theskills, the safeguards and the
use cases for your business.
That could then form the kernel ofan AI strategy for the future,
because the AI that we use now willbe the worst it will ever use,
because it will very quickly getbetter and better and better.
So an AI strategy for today.
If you were to waste your timetrying to create the perfect
(23:38):
document that is your AI strategywill be out of date in a couple of
years, if not sooner.
So thinking about how AIand generative AI particularly
is going to affect your businessmodel is a bit of a
fool's game unless you're in thegenerative AI business, or maybe
the software business for the restof us.
Start by trying it out.
(23:58):
Get your staff comfortable withprompting, and start to bring
in the kind of data science.
Skills that you'll need to thendifferentiate with AI.
That's what I think mostbusinesses should be doing.
That's what I'm advising our clientsto do. Unless you've already got
very, very deep data science skills.
Can we go back on control trial?
And, um, what I'm like for a pilot,but I'd be interested to hear
(24:20):
thoughts from both of you. What doesa good control trial look like?
So I think there's a bit ofa balance here because as Chris
said, you know, you need people toexperiment. You don't want to be too
prescriptive about what you wantthem to do, but you do
need to be able tomeasure what they're doing.
You do need some feedback.
(24:40):
You've got to create forums wherebypeople can share their experiences,
where they can provide feedback,where they can evaluate the
functionality and share what'sgood and what's not good, and
work together and build on that.
If you just set up a trial and say,you know, see you in a year,
you're not going to get any valuefrom it.
And you do need someone tocoordinate that, don't you?
(25:01):
We've got a full time digitalcommunity coordinator, for instance.
Even he is snowed with trying tocoordinate all these different
professional workers and, you know,get what works and what doesn't and
get them to, you know, create theirtips and tricks and so forth.
I think also, you've got to involverisk in compliance from the start in
a trial, don't you.
Mhm. Um, you know you've got to havesome guardrails
(25:22):
is the kind of word that everyone'stalking about which simply mean some
do's and don'ts.
Don't use personal data.
Don't use it to create client orcustomer facing outputs for now
until we better understand some ofthe risks around it.
But do use it for anythinginternally that is associated
with research or meeting prep,or things like coding and,
(25:43):
uh, Excel macros and the like, whichwe've mentioned, which are much less
risky because codeis not something that's covered by
intellectual property.
As such, it either works or itdoesn't.
So, um, I think havingthose guardrails that people need to
understand is a key part of a trial.
There's a cultural point to, isn'tthere, like the nature
(26:03):
of a lot of people in the accountingprofession is they see this this
tool and they want to break it.
They'll ask it a question like, tellme how this tax legislation works.
Tell me how this IFRS applied andwhen it doesn't work, sort of sit
back and say, oh, we're safe.
It doesn't work.
Yeah. But actually part ofthe culture has to be to take
all the narrative and all theconversation around that testing and
(26:24):
make it constructive, whichtakes time.
People need to feel safe.
There needs to be some joy in it,doesn't there?
I don't know, that's how I feel.
But be curious how you feel Sarahabout the culture.
How do you get people playing withit that is constructive?
Yeah, and it sort of plays backto the point about the digital
community and havingpeople who feel part of
(26:47):
an environment where they feel safeto do that experimentation,
where they've got support if theyneed it, because, um,
sometimes just a little bitof a pointer in the right direction
can make a really big difference forwhatever it is somebody trying to
do. So, you know, havingthose collaboration helps
that community feel,um, but I think, you know,
(27:09):
on a broader level around theculture, um, all
of this does go hand in hand withbasic levels of data literacy,
which are part of agood data strategy, is, you know,
it can't just be about thetechnology, the tools, the data.
It needs to be about how peopleinteract with those things, how
(27:29):
they feel comfortable and getconfidence in the inputs
and the outputs, howthey feel empowered to challenge,
to to question when they thinksomething might not be right.
And all of that comes back to basicdata literacy and just asking
questions of where did that datacome from? How did it come up with
that answer?
Am I sure it's the right answer?
(27:50):
How do I cross-check?
And that's just a big educationpiece and it doesn't happen
overnight. You have to kind of keepgoing with that.
I think also peopleare used to the world of software
either working or not working, butwith generative AI, the shades
of grey, it worksvery well most of the time, but
sometimes it hallucinates andoccasionally it just gets things a
(28:11):
bit muddled up or whatever.
But human beings are fallible too,and so I think that it's
got the reason it reached aninflection point.
I think after years of developmentof, um, language models
and so forth, was because itsuddenly became
as good in many scenarios as a humanbeing, because we're all fallible,
we can put errors into ourinto our written outputs and so
(28:33):
forth.
And so it's it's not ablack or white, it works or it
doesn't.
Largely it's good enough for many,many use cases.
So we shouldn't be waiting forperfection.
We should be using it now.
But with the right review around it,you know, it's an assistant isn't
it? It's assistive technology.
And as long as you review what itproduces, we should be using it
(28:55):
now.
So a year ago, broadly,no one had heard of Gen AI.
I know people had, but generalrecognition was very, very low.
I think now it's fair to say mostpeople have heard of it and played
with it. The two questionswhere will we be in 5 or 10 years
time?
Who wants to tackle five years andwho wants to tackle ten years?
Uh, that's sohard, isn't it? Because it's
(29:18):
evolving at such a pace thatI think it's pretty much impossible
to predict specifically what it'llbe capable of in our own profession.
There's a lot of assumptions aroundthe types of compliance
work, the types of advisory workthat a human being is
essential for.
Now, I still think that human beingswill be essential for most of that
(29:38):
work in the future, but in muchsmaller ways.
Um, and so I thinkthat the ability
of generative AI to incorporatea lot more data from the real world
that go into all of ourthought processes when we as
human beings are producing, um,written or visual outputs
(30:00):
can be incorporated into the modelsin the future, and it will be able
to do a lot of things that humanbeings currently assume that they
are. Um, you know, thatthe masters of for now.
So, I mean, if I'm going to make aprediction, is that a lot of the
current assumptions around what ahuman being is needed for will be
gone.
But in terms of specifics,if only I knew Ben!
(30:23):
You mentioned it.
A lot will change.
But is it too early to reflectdeeply on the impact of this
technology because it's so muchunknown?
I don't think it is.
I don't think it is. I mean, we youknow, there's a lot of examples of
generative AI in our daily livesthat a lot of people don't even
realise is, is AI enabled,you know, the Siri and the, um,
Amazon Alexa and all those kind ofthings, which, again, are getting
(30:44):
better and better all the time.
Um, so I think it's aboutgenerative AI being much
more woven into the fabric ofour daily lives.
You know, in cars, in,uh, bus stations and train
stations, in airports as well asin the working environments.
I think it's going to be a case ofit will augment most of what we
(31:06):
do in our daily lives, personallyand professionally.
That's what we need tostart thinking about now.
And obviously, the Bletchley Parksummit recently was a part of that
to begin to think about the privacy,um, and regulatory aspects
of it.
It's definitely not too soon to bethinking about that.
We've got to be thinking about an AIaugmented, you know, future
(31:27):
of humanity candidly.
But to bring it back to the businessworld.
All businesses are going to be AIenhanced, or most businesses already
are, even if they don't realise it,but becoming more explicitly so in
the next few years.
Yeah, and I think to build onthat, you know, the first
phase, I'm not going to put years onit, but just the first phase is
definitely about embeddinga different way of working and a
(31:49):
different way of thinking, uh, youknow, across personal lives
and professional lives.
Um, I think what we'll see,let's say, within the 5 to 10 year
window is a lot more proprietary,um, AI, generative
AI and AI models whereorganisations invest
in building something that is very,very personal to them and their
(32:11):
workforce.
But that isn't that can't happenquickly. It's not going to happen
overnight. It will take a lot ofinvestment. It's quite complicated.
So I would counsel that,organisations don't
rush into that.
Start thinking about what thoseuses, use cases might be,
what would be the underlying datathat's needed.
Get your workforce ready andused to working with this kind
(32:33):
of technology, and then investin the right use case.
And it might just be one.
Make sure it's the one that's goingto bring the most value.
Thank you so much.
Um, Chris and Sarah, um,if you would like to learn more
about Genrative AI and someinsights, please take a look at our
website and we'll put a link to thatin the show notes.
And please look out for furtherepisodes of The Loop, where we'll
(32:54):
continue to investigate the impactof this technology on businesses.
And thank you very much forlistening.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.