- il y a 2 jours
The Human Algorithm: Leading Work and Talent in the AI Age
Catégorie
🤖
TechnologieTranscription
00:00Moderate this excellent panel.
00:01We have such experts to talk to.
00:03So, starting off talking about AI and work,
00:08one of the first questions that you people hear is,
00:11wait, what does this mean for my job?
00:14And there have been some very scary predictions
00:17about the impact on all of our jobs
00:20in basically replacing humans.
00:23And then there have been some other predictions
00:25that, no, this is going to be an incredibly exciting
00:28job-changing technology
00:31that will create lots of new jobs.
00:33So, I cannot imagine a better panel to talk about this.
00:37So, Jim, why don't you start and give us a sense?
00:40Are the predictions, the scary predictions,
00:43are they overdone?
00:44No, I think every time we've introduced
00:47new technology in history,
00:48we have been able to increase the opportunity
00:52and then deal with the change.
00:54I think this time, I believe the opportunity
00:56of adding AI, not so much to the consumer world,
01:01but to the real world, as I call it,
01:03the reinvention of energy systems,
01:06of healthcare systems,
01:07of transportation systems,
01:08of all the physical infrastructure,
01:11when we make that more intelligent,
01:13we can make it more sustainable,
01:15we can make it significantly cheaper
01:17and more relevant.
01:20And I think that opportunity itself,
01:22I believe, is the biggest opportunity
01:24in my lifetime.
01:26And therefore, I look at it as a benefit.
01:30Now, the transition is hard.
01:32And the main problem in the transition
01:35is it's going to go very fast.
01:38And so, the limiting factor will be
01:41our ability to inspire human beings
01:44for different roles,
01:46because otherwise,
01:48they will become frustrated
01:49over the speed of change
01:50and the loss of opportunity for them.
01:53Jonas?
01:54Yeah, just like Jim says,
01:56I think this is probably
01:57one of the seminal moments
01:59of a massive opportunity and change,
02:02as we have seen
02:03under, you know,
02:05in different times
02:06as, you know,
02:08through human history.
02:10But I think what we know today
02:13is that technology is moving
02:15at a faster pace
02:16than we would have expected,
02:17yet adoption is what's holding us back
02:22to see the results.
02:23So much as we think about
02:25how industrial automation took time,
02:28not so much because of technology,
02:30but because of the changing work process,
02:32the different skills needed,
02:34how humans want to adopt
02:36and how they want to work
02:37versus what technology could do,
02:39I think the transition time
02:41will actually be quite significant
02:44because it is a lot of change management.
02:46But the overall promise remains
02:48tremendous and very, very exciting.
02:53Ard?
02:54I'm very much aligned.
02:55We see a lot of noise,
02:56a lot of prediction,
02:58and we are not in the business
03:00of, like, formulating those predictions,
03:02especially not being a tech company.
03:04So we have our eyes wide open.
03:07We recognize there is an impact,
03:09but we also know that in the past,
03:12you know,
03:12and specifically coming from the Silicon Valley,
03:15we've seen some very extreme point of view
03:17coming from working from home,
03:20for example,
03:21where we felt super old-fashioned,
03:22not doing, like, full remotes at some stage,
03:25and then things rebalanced.
03:26So what we try to do at LVMH
03:28is to have a very balanced approach,
03:31recognize the change,
03:32the pace of the change,
03:34because it's coming very fast,
03:36and to be very pragmatic.
03:38And to give you an optimistic view,
03:42I recently interviewed a Harvard professor
03:44named David Deming,
03:45who has studied technology transitions before,
03:48and he says that one interesting thing
03:51is that almost always there is a prediction
03:54that there will be job destruction
03:57and nobody will have a job anymore,
04:00and that actually every time
04:02way more new jobs are created.
04:04And one of the distinctions that he draws
04:06is that there are really two kinds of technologies.
04:10There are technologies that replace jobs,
04:12like tractors on farms,
04:14and then there are technologies that change jobs,
04:18like spreadsheets or word processors.
04:20And in your experience,
04:22you have such a wide range of experience.
04:24What are you seeing so far?
04:26Is AI changing people's jobs,
04:29or is it actually eliminating jobs
04:31and creating new ones?
04:33Aud, do you want to start?
04:34I think at that stage,
04:36what we see in LVMH,
04:37it's augmenting some of the jobs,
04:40because we're also an industry of craftsmanship.
04:43You know, we do things by hand.
04:45We have atelier,
04:47so that part is not being touched.
04:48There are some other jobs,
04:50you know,
04:50that'd be like sales planning,
04:52operations, HR,
04:53where we see that we are making faster,
04:55more clever decisions with AI,
04:58but it's not replacing.
05:00That's said,
05:00there is also a scale of change,
05:02and we recognize
05:03that we are probably at the beginning.
05:05So we will also keep watching what's coming up.
05:09Yeah, we estimate that a lot of this
05:12will be augmenting human capabilities
05:14as opposed to replacing them,
05:17but it will require then different skill sets
05:20to take advantage of this new technology.
05:22So it's not, you know,
05:23we can all continue doing what we're doing,
05:25and somebody is going to be on our shoulders
05:26telling us what to do.
05:27We also have to work in a different way.
05:30You know,
05:30if we're referring to Gen AI,
05:32Gen AI today is very good
05:34with content, conversation, and coding.
05:37So the areas where you're seeing
05:39a more major impact in terms of jobs
05:41clearly is in coding and programmers.
05:45So programming can be done
05:46with less people and achieve more,
05:50and it's why you see a prevalence of AI
05:53a lot within R&D
05:54and within customer service centers as well.
05:57So those are the areas
05:58that are beginning to,
05:59where we see at scale,
06:00an augmentation that is visible.
06:04But for the rest,
06:05just as Maud says,
06:06you know,
06:06we're at the very early stage
06:08of this evolution.
06:09So it's difficult to see
06:11at a macro level
06:12the impact on labor markets,
06:14with the exception maybe of programmers,
06:16where the demand has come down
06:18quite significantly
06:19over the last 18 months.
06:22Well, I think that reference
06:24to the generative AI
06:26is an important distinction.
06:28We've seen the revolution
06:30come through the large language models,
06:32which, you know,
06:33seem very human
06:34because they were taught
06:35human language.
06:37And so their interaction
06:39is inspiring us.
06:40And I think there will be
06:42a huge category
06:43around productivity,
06:45of basically taking over
06:47the things we do
06:48which are not very productive,
06:50like all the emails we write.
06:52You know, nowadays,
06:53we often enhance the email
06:55with AI and who knows,
06:57one day AI will write emails
06:58to AI
06:59and we can do other things
07:00that are more productive
07:01than sit and write our emails
07:03or file our travel expense reports.
07:08So if you look
07:09at the white-collar worker today,
07:11we spend a lot of time
07:13on things that don't add much value.
07:16And if we can add
07:18these language models,
07:19I think we can unleash
07:20that human potential
07:21in new ways.
07:23But if I look at the way
07:26C3 AI is installing
07:28not just language models,
07:30but real data AI
07:33in specific domain areas,
07:37we see two more categories,
07:39not just productivity
07:40from language models.
07:41We see companies
07:43becoming more predictive,
07:45predictive in predicting
07:47failures of assets,
07:49in predicting issues
07:51in supply chains,
07:52and in a world
07:53of dramatic disturbance,
07:56being more predictive
07:57and being ready
07:58for a disruption,
08:00of course,
08:01means dramatic improvement
08:03of business outcomes
08:04and, you know,
08:05competitiveness.
08:05So there's a whole category
08:07around becoming
08:08more predictive.
08:09I still imagine
08:10the boardroom
08:11where we don't talk
08:12about last quarter,
08:13but next quarter
08:14because we are becoming
08:15predictive
08:16in our way
08:17of dealing
08:17and leading.
08:18and I think
08:20there's a third category
08:21now on the horizon
08:22which is about innovation,
08:25using these deep
08:28data-driven AI models
08:30to enhance
08:32our innovation capability.
08:34Roland Bush,
08:35the CEO of Siemens,
08:36which I chair,
08:38will be on stage today.
08:39NVIDIA Jensen
08:40was on stage earlier today.
08:42And the collaboration
08:43around that
08:43is really to imagine
08:45the world
08:46of designing products.
08:47Imagine a generative AI
08:50in the industrial world
08:51where you say,
08:52please help me design
08:53a two-seat sports car
08:55that is electric,
08:56goes 700 kilometers
08:58on one charge
08:59and looks a little bit better
09:00than a Porsche 911.
09:02You can imagine
09:03how this changes
09:05the opportunity
09:06for innovation
09:07where the human being
09:09is enhanced
09:10and can jump over
09:13easy phases
09:14and add the brilliance
09:16of human,
09:17which I think
09:18only humans can do.
09:19But wow,
09:20if we could speed
09:21the innovation process up.
09:22And this is going
09:23way beyond language models.
09:25You can't run your company
09:26based on the most
09:28next likely word
09:29in a sentence.
09:30You need AI
09:31that is deep
09:32and understanding
09:33the data of your company,
09:35the specifics
09:36of your products
09:37and the design
09:38of your hardware.
09:40And then you can create
09:41these dramatic improvements,
09:43which I see
09:44is coming right now
09:45with the efforts
09:46we do with C3
09:47as an example.
09:48And so one thing
09:49we all seem to agree
09:51is that this is a time
09:52of very great change
09:54and really the question
09:56is how fast
09:57does it happen?
09:58It's difficult,
09:59and I know this
10:00from first person managing,
10:03it's difficult
10:03to manage change
10:04at companies
10:05because there is
10:07often anxiety,
10:08there's excitement,
10:09people don't know
10:10how their jobs
10:10are going to change.
10:11So what are some examples
10:13of how you are managing
10:15change
10:16and what you're doing
10:17that is working
10:18within your organization
10:20with AI?
10:22I think we've been
10:23managing change forever,
10:24and the last 10 years
10:26we've seen
10:27an acceleration of change.
10:29You know,
10:3010 years ago
10:30we were having
10:31conversation about
10:32do we need to sell
10:33some of our handbags
10:34on e-commerce websites?
10:36Some of our maisons
10:38said yes,
10:38said no,
10:39there was some hesitation
10:40but we've built
10:41infrastructure
10:42for that change management
10:43and we had
10:44some extreme position.
10:46You know,
10:46in 2020
10:47we thought like
10:48most of the retail estate
10:49would probably close
10:50and everything
10:51would be sold online
10:52and then we went back
10:53to that prediction
10:54and what we see today
10:55is probably 10%
10:57of the business
10:57of the LVMH group
10:59that is done online.
11:01That has given us
11:03the team,
11:04the structure
11:04and the tech
11:06structures
11:07to really lead
11:08that new change
11:10that is coming.
11:11So it's not
11:12where the challenges
11:13for us are lying.
11:14It's really
11:14on that human journey
11:16that is AI
11:18and how we'll adopt
11:19it in the jobs.
11:21I think the additional
11:22challenge for LVMH
11:23is driving it
11:24for a decentralized
11:26organization
11:26because to implement
11:28AI,
11:29Gen AI,
11:29you need scale
11:30and you need processes.
11:32we don't always
11:33have the scales
11:34because of the
11:35decentralization
11:36so we need to go
11:38and we have pilots
11:39in the different
11:41maisons
11:41then we can scale back.
11:44I think those pilots
11:45are really giving us
11:46an opportunity
11:46to train
11:47and to learn
11:48in the teams
11:49in the different
11:50functions
11:51to take that learning
11:52and then to be able
11:53to scale back
11:54the learning
11:55in the different
11:56functions
11:57and divisions.
11:58So we have
11:58a very pragmatic
11:59view on it
12:01and then of course
12:02there are some modules
12:03like the Gen AI systems
12:05Maya that we are using
12:06and it's for the
12:07215,000 employees
12:09in the group
12:10in one go.
12:11There was also
12:12a lot of discussion
12:13internally
12:13on who's supposed
12:14to lead that transformation.
12:16Is it tech?
12:17Is it digital?
12:17Is it HR?
12:18Of course I'm going
12:19to say it's HR.
12:20It's a human transformation
12:22but we've seen
12:23examples recently
12:24of Moderna
12:25for example
12:26merging their tech
12:28and HR teams
12:30and you're like
12:30wow
12:31that's probably
12:32totally opposite
12:33like machine
12:34and people
12:35and they work together.
12:36It's blurring
12:37the line
12:38in the organization
12:39and what we see
12:41is like
12:42if we manage
12:43to do it
12:43dramatically
12:44job by job
12:45also having
12:45probably part
12:46of the business
12:47where we put limits
12:48where we look
12:49at responsibility
12:50at ethics
12:51around AI
12:52it's how we will
12:53curve you know
12:54that transformation
12:55for the years
12:56to go
12:56but we see it
12:58in our jobs
12:58it's very like
13:00real in the business
13:02now we need
13:03to find the scale
13:04we finish the pilots
13:05in the different businesses
13:07to see where it takes us
13:09in the five
13:10ten years to go.
13:12and very similar
13:13to what Moderna
13:14is talking about
13:15we have a human
13:16first and digital
13:17always approach
13:18within the company
13:19and you know
13:20some of the things
13:21in addition to
13:22very similar things
13:22that we're doing
13:23we're trying to
13:24you know
13:25to start
13:25we have to be
13:26very clear
13:27about the ethics
13:28around AI
13:28you know
13:29what we would accept
13:30as an organization
13:31what our people
13:32could do
13:32and what they
13:32couldn't do
13:33because we deal
13:34with you know
13:35millions
13:35of people's data
13:37so we have to be clear
13:38about what's an acceptable use
13:40an ethical use
13:41versus what is not
13:43and I would say
13:44we're also seeing
13:45and emphasizing
13:47you know
13:48a bit of the
13:48generational differences
13:49that we're seeing
13:50many of our younger workforce
13:52are almost native users
13:54and you know
13:55when we ask them
13:55how do you use AI
13:56sometimes they don't
13:57understand the question
13:59because it's embedded
14:00in how they live
14:01their lives
14:02and how they just
14:03help themselves
14:03doing things that work
14:06but middle
14:07and senior management
14:08who have not
14:09grown up with AI
14:10are actually struggling
14:12much more
14:12with the adoption
14:14and you know
14:15if you take myself
14:16as an example
14:16I tended to think
14:18about AI
14:18as a nephew
14:19that my sister
14:20wanted me to hire
14:21so nepotism
14:23at its worst
14:24with a nephew
14:25that I was wondering
14:26when he was going
14:27to be screwing up
14:28and then he screws up
14:29and I say
14:30yes
14:30he didn't know
14:32how to do it
14:32I knew it
14:33I knew he wasn't
14:34going to do it
14:34but instead
14:35shifting to this notion
14:37of no no
14:37AI is like a genius
14:39I've been trying
14:39to attract and hire
14:41for three years
14:41it's been impossible
14:42but now
14:43she is in the company
14:45and we are
14:46good to go
14:47and everything
14:48that comes out
14:49might be wrong
14:50sometimes
14:51but largely
14:51is going to be
14:52very positive
14:53and that's a mental
14:54change management
14:56that I think
14:56we see
14:57is more difficult
14:58with senior managers
15:00who have a lot
15:01of experience
15:01they know
15:02what you start with
15:03and they know
15:04what you end with
15:04and they know
15:05what's worked
15:05for 20 years
15:06so getting
15:07that view
15:08to change
15:09is a much
15:10bigger change
15:11exercise
15:11actually
15:12than what it is
15:13with the younger
15:14generation
15:15in our workforce
15:17yeah
15:17so I
15:18recently
15:20compared
15:20AI
15:21and the
15:23let's say
15:23deployment of AI
15:24with another
15:25technology
15:26that profoundly
15:27changed our
15:28society
15:28and everything
15:29in human interaction
15:32which was
15:32the printing press
15:34before the printing press
15:37knowledge was centralized
15:39around few people
15:40who could afford
15:41to spend the time
15:43and hand write a book
15:45and give it to someone else
15:46you couldn't distribute
15:48at large knowledge
15:49and with the printed press
15:50you basically had
15:51the democratization
15:52of knowledge
15:53and ideas
15:53so everyone
15:54could participate
15:55I think AI
15:56is the same
15:57but this is the
15:58democratization
15:59of intelligence
16:00of super intelligence
16:01and the biggest
16:03difference this time
16:04in terms of deployment
16:05is that it took
16:06roughly 400 years
16:07to teach
16:08the population
16:09at large
16:09to read and write
16:10so that they could
16:11take advantage
16:12of this technology
16:13well this time
16:14AI speaks human language
16:16it's taught
16:17with human interaction
16:18so there is
16:20very little
16:21training
16:21the training
16:22is about
16:23understanding
16:24like was mentioned
16:26already
16:26how can I
16:27use this technology
16:29to become better
16:30now you take
16:31a look at that
16:32and you imagine
16:33what it does
16:34to an organization
16:34where in traditional
16:36organizations
16:37we would spend
16:38a lot of time
16:39training newcomers
16:40into the organization
16:41so they understand
16:42how we do things
16:43how machines work
16:44how processes work
16:45and suddenly you have
16:46the intelligence
16:48of in the case
16:49of Siemens
16:51176 years
16:52of history
16:52at a fingertip
16:54for every single
16:55employee
16:55and every single
16:56customer
16:57of these pieces
16:59of equipment
17:00instantly available
17:02imagine what that
17:04does to the human
17:05potential
17:05it's almost like
17:06we spent 200 years
17:07teaching human beings
17:09to act like machines
17:10so we can
17:11predict outcomes
17:12of work
17:13and now we can
17:14unleash that work
17:15and have human beings
17:17be human
17:18and creative
17:19and create value
17:21at a level
17:21we have never
17:22seen before
17:23now the deployment
17:24of that is difficult
17:25because you want
17:26to decentralize
17:27the access to AI
17:29and even the
17:30development of AI
17:31but you want to
17:32centralize
17:33the technology
17:34choices
17:35you want to
17:36choose one
17:37platform
17:37otherwise you get
17:39chaos
17:39and no scale
17:40and I think
17:41this is the dilemma
17:42you have
17:43in the deployment
17:44the decentral use
17:46but the centralized
17:47governance
17:48otherwise
17:49you lose control
17:51of one of the
17:51most powerful
17:52technologies ever
17:53deployed
17:54in IT
17:55and business
17:56and you're all
17:57talking about
17:58managing change
17:59in the organization
18:00and Jonas
18:00you talked about
18:01it sounds like
18:02most of the
18:04intelligent usage
18:05of AI
18:06came from the
18:06younger people
18:07who knew how
18:08to do it
18:08and senior management
18:10like me
18:10scared
18:11maybe we will
18:12get hurt
18:12maybe it will
18:13be wrong
18:13what is the best
18:15is there a best
18:16way for a big
18:17company to
18:19approach it
18:20is it a top
18:21down approach
18:22to adopting AI
18:23or is it
18:24seeing what is
18:25happening within
18:26the organization
18:27and encouraging
18:28that
18:29I'm guessing
18:30it's a combination
18:30of both
18:34bottom up
18:35experimentation
18:36and top down
18:38governance
18:38and encouragement
18:40of daily
18:41use
18:42and I think
18:43that's sort
18:44of the trick
18:44we're not used
18:45to having a tool
18:46like this
18:46intelligence
18:47on tap
18:47just like
18:48you know
18:49Jim is talking
18:49about
18:50so we have
18:50to rethink
18:52how work
18:53gets done
18:54and you know
18:55that's something
18:56that is a change
18:56and all change
18:57is difficult
18:58most people
18:59don't want to
19:00change
19:00you know
19:01unless they have
19:01to
19:02and this is
19:02one of those
19:03where the
19:03opportunity is so
19:04big we have
19:05to help
19:05everybody
19:06make this
19:08transition
19:09what we've seen
19:10as well
19:10just to add
19:11you know
19:11at individual level
19:12and I know
19:13it's in many
19:13organizations
19:14is that shadow
19:15use of AI
19:16and people
19:17nearly feeling
19:17guilty
19:18and doing it
19:19you know
19:19when they work
19:20from home
19:20or remotely
19:21and not sharing
19:22you know
19:23what they were
19:23finding
19:23because they were
19:24finding those
19:25level of efficiency
19:26what do you do
19:27of those
19:28efficiency gain
19:29so it's also
19:30having that
19:31governance
19:32that bottom up
19:33and giving
19:34the permission
19:34to share
19:35on what they've
19:36done
19:37that others
19:37know that
19:38they are using
19:39it
19:40we have a very
19:41strict governance
19:41at LVMH
19:42both in terms
19:43of ethic
19:44and compliance
19:45but also
19:46around confidentiality
19:47and making sure
19:48that we use
19:49our own system
19:50because we know
19:50that in terms
19:51of confidentiality
19:52bridge
19:53you know
19:53there are huge
19:54risks
19:54for the organization
19:55in terms
19:56of losing
19:56some of that
19:57knowledge
19:57so I would say
19:58that having
19:59that one-to-one
20:00and managing
20:01to train
20:01people one-by-one
20:03in those large
20:03organizations
20:04is also
20:05one of the
20:05challenges
20:06and I think
20:07what's fantastic
20:07with AI
20:08is that AI
20:09is actually
20:09helping you
20:10train
20:10people at scale
20:12so you know
20:12it's the reverse
20:13benefit
20:13of using it
20:15Jim
20:16you know
20:17as I mentioned
20:18I think
20:19so I have
20:20some experience
20:21with this
20:21since 10 years
20:22roughly now
20:23and the first
20:24projects
20:25of AI
20:26in the organizations
20:27where I was
20:28involved
20:29was very
20:29decentralized
20:30you would hire
20:31some data scientists
20:32who would speak
20:33languages
20:33that nobody
20:34would understand
20:35and they would
20:36do their pilots
20:36the pilots
20:37would be successful
20:38but they never
20:39scaled beyond
20:40the pilot
20:41then you
20:42centralize this
20:42and say
20:43no no no
20:43this doesn't
20:44work
20:44we need to
20:44control
20:45the use
20:46of AI
20:46the data
20:47everything
20:48and now
20:49you get
20:49you know
20:49perfect scale
20:50but no usage
20:52and you need
20:53a combination
20:54of the two
20:55I think you
20:55need to be
20:56almost like
20:57a dictator
20:58on choosing
20:58one technology
21:00stack
21:00one strategy
21:02around data
21:02so that you
21:03get the base
21:04in order
21:05and then
21:06you decentralize
21:07the use
21:08of that stack
21:09of that
21:09technology
21:10so that you
21:11can have
21:12the maximum
21:12speed
21:13and the focus
21:14not on
21:14technology
21:15choice
21:15and discussions
21:16whether this
21:17tool or that
21:18tool is better
21:18just use the
21:20technology
21:20and dramatically
21:21accelerate the
21:22adoption
21:22that can
21:23happen today
21:24in modern
21:25AI platforms
21:26for the enterprise
21:27where you
21:28don't have to
21:29be a data
21:29scientist
21:30you just have
21:31to understand
21:32business
21:32you don't even
21:33have to program
21:34anything
21:34the system
21:36understands
21:36the data
21:37it understands
21:38your role
21:39and will
21:39help and
21:40guide you
21:40develop
21:41predictive
21:41and incredible
21:42value
21:43with very
21:44little effort
21:45and certainly
21:45no special
21:47expertise
21:47because the
21:48programming
21:49has gone away
21:50in the background
21:51and these AI
21:52systems will help
21:53you do that
21:54so I think
21:54that's the
21:55combination
21:56centralized
21:57governance
21:58is almost like
21:58dictatorship
21:59on technology
22:00and then
22:01decentralized
22:01and accelerated
22:02use
22:03where everyone
22:04becomes an AI
22:05wizard
22:05because everyone
22:06needs access
22:07to this
22:08technology
22:09and what are
22:11the best practices
22:12for training
22:13because I think
22:14within organizations
22:15most people
22:16want to do
22:16a good job
22:17they want
22:17their boss
22:18to think
22:18wow this is
22:19a good person
22:20and they're
22:20not being
22:21irresponsible
22:21and so forth
22:23and yet
22:24when you have
22:25not used
22:25AI yourself
22:26generative AI
22:27in particular
22:28often it's
22:29scarier
22:30than when you
22:31actually use it
22:32you learn
22:32what it can do
22:33what it can't do
22:34how do you
22:36train
22:36your organization
22:38to embrace
22:38that
22:39without using
22:40it recklessly
22:42I think
22:43first we had
22:44some
22:44training
22:46that was like
22:47really deployed
22:48at group level
22:49and you know
22:50when it comes
22:50to dictatorship
22:51in technology
22:52at LVMH
22:53we have
22:53the captain
22:54that is making
22:55sure that
22:56everybody
22:56is using
22:57the same one
22:58and we
23:00reached everybody
23:01we had
23:02a first level
23:03then a second
23:03level
23:04and then
23:05what we want
23:05is really
23:06people to use
23:06it in the job
23:07because otherwise
23:08you don't really
23:09reach huge
23:10transformation
23:10I think in
23:12terms of
23:12efficiency
23:12of course
23:13people become
23:13more efficient
23:14on the job
23:15but if you
23:15look at
23:16entire processes
23:17it needs
23:18that team
23:19are working
23:19on it
23:20different
23:20expertise
23:21different
23:22methi
23:22are joining
23:23so that's
23:24what we are
23:24doing
23:24through the
23:25pilots
23:25and they
23:26are learning
23:27by doing
23:28you know
23:28and we know
23:28that 70%
23:29of learning
23:30happens
23:30on the job
23:31and I truly
23:31believe it
23:32the same
23:33with AI
23:33and then
23:34once we have
23:35those learning
23:36we share
23:36at group level
23:37and we go
23:38very fast
23:38at scaling
23:40much the same
23:41we use
23:42we try to
23:43govern
23:43the tools
23:44that we're
23:44using
23:45learning
23:45by doing
23:46and I
23:47would say
23:48you know
23:49most jobs
23:50contain
23:5015 to
23:5120 tasks
23:52we try
23:53and get
23:54people to
23:54use it
23:55for a
23:55specific
23:56task
23:56and see
23:57the benefits
23:58and encourage
23:59the use
24:00and then
24:01we can
24:01combine it
24:02with two
24:02or three
24:02other tasks
24:03because you
24:04also have to
24:05trust the
24:06outcomes
24:06you have to
24:07trust what
24:07it is that
24:08you're doing
24:08so we're
24:09trying this
24:10on a step
24:10by step
24:11approach
24:11clearly
24:11encouraging
24:12you know
24:13learnability
24:14within the
24:14organization
24:15because ultimately
24:16in an environment
24:17like this
24:17a culture
24:19of curiosity
24:19and encouraging
24:21trying
24:21is going to
24:22be very
24:23important
24:23as things
24:24will change
24:24all the time
24:26there's not
24:26going to be
24:26a fixed
24:27training
24:27regimen
24:28you will
24:29have to
24:29sort of
24:29encourage
24:30people
24:30to keep
24:31on learning
24:32as they
24:32evolve in
24:33their careers
24:33because the
24:34technological
24:34change is
24:35coming very
24:36very quickly
24:38so from
24:39a learning
24:40point of
24:40view
24:41I think
24:41the biggest
24:41risk we
24:42have is
24:43that people
24:43you know
24:46see this
24:47black box
24:48and they
24:49get a
24:49response
24:50to something
24:50and they
24:52don't
24:52understand
24:53the background
24:54for this
24:55response
24:55if that's
24:57the way
24:57we teach
24:58human beings
24:59to use
24:59this
24:59technology
25:00it will
25:01soon
25:01go out
25:02of
25:02control
25:02and I
25:03can use
25:04it like
25:05that
25:05for creating
25:05a poem
25:06for my
25:07son or
25:07my wife
25:08but I
25:09can't use
25:09it like
25:10that to
25:10run a
25:10company
25:11I need
25:11no
25:12hallucinations
25:13to run
25:13my company
25:14so for
25:15me
25:15the education
25:16is about
25:17first of
25:18all
25:18a big
25:19principle
25:19around
25:20explainability
25:21I am
25:22a big
25:22believer
25:23that AI
25:24systems
25:24used in
25:25the
25:25enterprise
25:25need to
25:26explain
25:27why they
25:27came to
25:28this
25:28conclusion
25:28they need
25:29to be able
25:30to tell
25:30you what
25:31is the
25:31likelihood
25:31that this
25:32is true
25:32they need
25:33to tell
25:33you what
25:34are the
25:34sources
25:34that
25:35created
25:35this
25:36response
25:36and they
25:37need to
25:37create
25:38an
25:38opportunity
25:38for
25:39interaction
25:39so that
25:40you can
25:41investigate
25:41you can
25:42ask
25:42again
25:42you can
25:43go
25:43deeper
25:43and this
25:44is for
25:45me a
25:45whole
25:46different
25:46level
25:46of generative
25:47AI
25:47than what
25:48we see
25:48today
25:49where
25:49we
25:49kind
25:50of
25:50throw
25:50a
25:50question
25:51and
25:51get
25:51an
25:51answer
25:52back
25:52and
25:52believe
25:52in
25:53it
25:53for
25:53the
25:53enterprise
25:54we
25:54need
25:54explainability
25:55and we
25:56need
25:56to
25:56train
25:57the
25:57human
25:57being
25:58that
25:58you
25:58are
25:59in
25:59control
25:59you
26:00are
26:00accountable
26:01for
26:01the
26:02outcomes
26:02and
26:02the
26:03conclusions
26:04not
26:05the
26:05AI
26:05and
26:06therefore
26:06I
26:06give
26:07you
26:07tools
26:07to
26:08interact
26:08with
26:08AI
26:09in
26:09meaningful
26:09ways
26:10that
26:10allow
26:10you
26:11to
26:11be
26:12in
26:12charge
26:12then
26:13I
26:13can
26:13empower
26:14human
26:18big
26:18mistakes
26:18and
26:19I
26:19will
26:19be
26:20disappointed
26:20so
26:21I
26:21think
26:21this
26:22is
26:22key
26:22in
26:22the
26:23training
26:23and
26:24the
26:24design
26:25of
26:25AI
26:26systems
26:26for
26:26the
26:26enterprise
26:27explainability
26:28and
26:29human
26:29interaction
26:30to
26:30ensure
26:31the
26:31human
26:31person
26:32is
26:32in
26:33charge
26:33so
26:34hopefully
26:35this
26:35is
26:35a
26:35fun
26:36question
26:36I
26:36was
26:37listening
26:37to
26:37a
26:38Silicon
26:38Valley
26:38venture
26:39capitalist
26:39on
26:40a
26:40podcast
26:40recently
26:41extremely
26:42excited
26:43about
26:43the
26:43potential
26:44for
26:44AI
26:44but
26:45also
26:46so
26:47optimistic
26:48in a
26:49way
26:49about
26:50its
26:50impact
26:50on
26:51the
26:51job
26:51market
26:51that
26:52he
26:52thinks
26:52that
26:52most
26:53jobs
26:53will
26:53no
26:54longer
26:54exist
26:54very
26:55quickly
26:55so
26:56he
26:56doesn't
26:57know
26:57what
26:57to
26:57tell
26:58his
26:58children
26:58to
26:59study
27:00in
27:00school
27:00because
27:01he
27:01wants
27:01them
27:01to
27:02be
27:02successful
27:02you
27:03are
27:04all
27:04extremely
27:05successful
27:06professionals
27:07many
27:07many
27:08people
27:08look to
27:08you
27:09for
27:09inspiration
27:09what
27:10is
27:11your
27:11advice
27:11to
27:12the
27:12next
27:12generation
27:13the
27:13generation
27:14in
27:14college
27:14and
27:15in
27:15grad
27:15school
27:16now
27:16in
27:17terms
27:17of
27:17what
27:17skills
27:18to
27:18learn
27:19and
27:20how
27:20much
27:21effort
27:21do
27:22you
27:22put
27:22into
27:22learning
27:23AI
27:23how
27:23valuable
27:24is
27:24that
27:24going
27:24to
27:24be
27:25what
27:25advice
27:25do
27:26you
27:26have
27:27so
27:27maybe
27:27I
27:28start
27:28first
27:29of all
27:29maybe
27:30I have
27:30two
27:30answers
27:30first
27:31of all
27:31I do
27:31believe
27:32that
27:32even
27:32if
27:32we
27:33have
27:33the
27:33most
27:33intelligent
27:34AI
27:34systems
27:35next
27:35to
27:35us
27:36we
27:36need
27:37to
27:37learn
27:37fundamentals
27:38it's
27:39a
27:39little
27:39bit
27:39like
27:39the
27:40calculator
27:40that
27:41yes
27:41you
27:41can
27:42have
27:42the
27:42calculator
27:42calculate
27:43for
27:43you
27:43but
27:43if
27:44you
27:44don't
27:44understand
27:44math
27:45you're
27:46not
27:46going
27:46to
27:46be
27:46a
27:46good
27:46user
27:47of
27:47a
27:47calculator
27:48and
27:48in
27:49the
27:49same
27:49way
27:49I
27:49think
27:49we
27:50need
27:50to
27:50learn
27:50the
27:51fundamentals
27:51of
27:52the
27:52various
27:53topics
27:53that
27:53are
27:54important
27:54everything
27:54from
27:55languages
27:55to
27:56math
27:56to
27:57specializations
27:58I
27:59think
27:59the
27:59new
28:00thing
28:00we
28:00need
28:00to
28:00teach
28:01young
28:01kids
28:02is
28:03it's
28:04almost
28:04like
28:04we
28:04need
28:04to
28:05reverse
28:05the
28:05school
28:06system
28:06today
28:07we
28:07teach
28:08kids
28:08how
28:08to
28:09answer
28:09questions
28:10the
28:11final
28:11exam
28:12is
28:12a
28:13bunch
28:13of
28:13questions
28:13that
28:14people
28:15need
28:15to
28:15answer
28:16and
28:17I
28:17think
28:17the
28:18new
28:18skill
28:18we
28:18need
28:19is
28:19to
28:19figure
28:19out
28:20what
28:20questions
28:21to
28:21ask
28:22because
28:23that
28:23becomes
28:24the
28:24key
28:25to
28:26unlock
28:26the
28:26potential
28:27of
28:27AI
28:28and
28:29I
28:29feel
28:29good
28:29about
28:30that
28:30because
28:30when
28:30I
28:31went
28:31from
28:31being
28:31a
28:31CEO
28:32or
28:32tech
28:32company
28:33to
28:33be
28:33the
28:33chairman
28:34my
28:35role
28:35went
28:35from
28:35answering
28:36questions
28:36to
28:37asking
28:37questions
28:38and
28:38I
28:38can
28:39assure
28:39you
28:39it's
28:40much
28:40more
28:41fun
28:41to
28:41be
28:41the
28:41one
28:42asking
28:42the
28:42questions
28:43not
28:44answering
28:44them
28:46Jonas
28:48well
28:49it's
28:49still
28:49fun
28:49answering
28:50the
28:50questions
28:50too
28:51but
28:51I
28:52really
28:53like
28:53the
28:53way
28:54you
28:54think
28:54about
28:54this
28:55connected
28:55to
28:55curiosity
28:56asking
28:56questions
28:57but
28:58in
28:58terms
28:58of
28:58advice
28:59really
29:00I
29:00think
29:00about
29:00it
29:01in
29:01the
29:01same
29:01way
29:01as
29:02I
29:03done
29:03for
29:03most
29:03of
29:03my
29:04career
29:05is
29:05what
29:06you're
29:06passionate
29:06about
29:07and
29:07what
29:07you're
29:07interested
29:07in
29:08is
29:09a
29:09much
29:09higher
29:09likelihood
29:10that
29:10you'll
29:10be
29:11successful
29:11in
29:11than
29:12the
29:12opposite
29:12of
29:13choosing
29:13a
29:13path
29:14that
29:14seems
29:15predetermined
29:15and
29:16and
29:16I
29:16look
29:16at
29:17the
29:17advice
29:17we
29:18heard
29:18and
29:19read
29:19about
29:19in
29:20the
29:20papers
29:2010
29:21years
29:21ago
29:21if
29:22you
29:22want
29:22to
29:22know
29:22where
29:23to
29:23go
29:23become
29:24a
29:24coding
29:25person
29:25people
29:26can
29:26be
29:26this
29:27and
29:27they
29:28can
29:28become
29:28coders
29:29and
29:29your
29:29future
29:30will
29:30be
29:30bright
29:30and
29:31forever
29:31rosy
29:32even
29:32if
29:32you
29:32don't
29:33like
29:33it
29:33it
29:33will
29:33be
29:33bright
29:34and
29:34rosy
29:34and
29:35here
29:35we
29:35are
29:3510
29:36years
29:36on
29:36and
29:37we
29:37report
29:37out
29:37that
29:38the
29:38only
29:38category
29:39where
29:39we
29:39can
29:39see
29:40a
29:40macro
29:40impact
29:40of
29:41technological
29:41change
29:42is
29:42coding
29:43so
29:44predicting
29:44what's
29:45going
29:45to be
29:45the
29:45next
29:46big
29:46job
29:46and
29:46what
29:47people
29:47are
29:47going
29:47to
29:47be
29:47doing
29:48I
29:48think
29:48is
29:48very
29:48difficult
29:49it's
29:49more
29:49a
29:50question
29:50of
29:50how
29:50so
29:51what
29:51are
29:51you
29:52passionate
29:52about
29:52and
29:53the
29:53curiosity
29:53we
29:54might
29:54have
29:54a
29:54whole
29:55school
29:56change
29:57a
29:57fundamental
29:58change
29:59in
29:59how
29:59we
29:59learn
29:59and
30:00being
30:00much
30:01better
30:01at
30:01asking
30:02the
30:08with
30:08the
30:08answer
30:08so
30:09what
30:09we
30:09question
30:10and
30:10what
30:10we
30:10do
30:11with
30:11the
30:11information
30:11will
30:12determine
30:12the
30:12outcome
30:13more
30:13than
30:14just
30:14the
30:14raw
30:15axis
30:15of
30:16intelligence
30:18I
30:19think
30:19what
30:19helps
30:20us
30:20you know
30:20asking
30:20those
30:21questions
30:21is
30:21also
30:22experience
30:22and
30:23if
30:23we
30:23were
30:24really
30:24to
30:24simplify
30:25like
30:25you
30:26know
30:26AI
30:26it
30:27would
30:27be
30:27like
30:27a
30:27human
30:28being
30:28that
30:28would
30:28have
30:29read
30:29millions
30:29of
30:30books
30:30and
30:31I
30:31think
30:31that
30:31is
30:32probably
30:32one
30:32of
30:32the
30:32critical
30:33skills
30:33when
30:34we
30:34look
30:34at
30:34the
30:35next
30:35generation
30:35for
30:36me
30:36is
30:36reading
30:36because
30:37reading
30:37is
30:38what
30:38helps
30:38you
30:38define
30:39your
30:39interest
30:40have
30:40your
30:41knowledge
30:41and
30:42really
30:42being
30:42able
30:43to
30:43have
30:43that
30:44critical
30:44spirit
30:45ask
30:46questions
30:46to
30:46AI
30:47understand
30:47how
30:47it's
30:48been
30:48made
30:48and
30:49know
30:49who
30:49you
30:49are
30:50so
30:50what
30:50we
30:51see
30:51at
30:51the
30:51moment
30:51with
30:52the
30:52speed
30:52of
30:52content
30:53with
30:53the
30:53fact
30:54that
30:54the
30:54content
30:54is
30:55not
30:55shorter
30:55has
30:56a
30:56huge
30:56impact
30:57on
30:57our
30:57brain
30:57and
30:58if
30:58we
30:58want
30:59human
30:59beings
30:59to
30:59remain
31:00like
31:00critical
31:00people
31:01and
31:02AI
31:02to
31:02be
31:03human
31:03led
31:03and
31:04that
31:04the
31:04reason
31:04why
31:05we
31:05build
31:05a chair
31:06we
31:06stand
31:07for
31:07at
31:07RVMH
31:08because
31:08we're
31:09a
31:09human
31:09led
31:10industry
31:10that's
31:10what
31:11we
31:11believe
31:11in
31:11I
31:12think
31:12reading
31:13is
31:13probably
31:13one
31:14of
31:14the
31:14best
31:14advice
31:15terrific
31:16and
31:16then
31:17last
31:17question
31:17which
31:18is
31:18that
31:18we
31:18have
31:19incredible
31:20executives
31:20in
31:21the
31:21audience
31:22wondering
31:23how
31:23do I
31:23handle
31:27wonderful
31:28guidance
31:28any
31:29last
31:30overarching
31:31pieces
31:32of
31:33advice
31:33on
31:34how
31:34people
31:35in
31:35business
31:36either
31:36junior
31:37or
31:37senior
31:38or
31:38middle
31:38how
31:39should
31:39we
31:39approach
31:40AI
31:41and
31:41anything
31:41to
31:42avoid
31:42any
31:43mistakes
31:44yes
31:45so
31:45I
31:45am
31:46often
31:46using
31:47a
31:47quote
31:48from
31:48Bernard
31:48Shaw
31:49who
31:49said
31:49most
31:50people
31:50see
31:50things
31:51the
31:51way
31:51they
31:51are
31:52and
31:52ask
31:52why
31:52I
31:54imagine
31:54things
31:54that
31:55never
31:55were
31:55and
31:56ask
31:56why
31:56not
31:56and
31:58so
31:58I
31:58think
31:58the
31:59one
31:59advice
31:59I
32:00have
32:00is
32:00to
32:00be
32:01imagining
32:02a
32:03future
32:03of
32:04AI
32:04instead
32:04of
32:05looking
32:05at
32:05the
32:05current
32:06my
32:06prediction
32:07on
32:07AI
32:07it's
32:08going
32:08to
32:08be
32:09everywhere
32:09on
32:10our
32:10phone
32:11in
32:11our
32:11glasses
32:11in
32:12our
32:12cars
32:13in
32:13everything
32:13we
32:14interact
32:14with
32:14I
32:15predict
32:16it
32:16will
32:16be
32:16super
32:17intelligent
32:17not
32:18a
32:18general
32:19language
32:20model
32:20specific
32:21domains
32:22where
32:23in
32:23each
32:23domain
32:24like
32:24protein
32:25folding
32:25or
32:25math
32:26or whatever
32:26it is
32:27just
32:27better
32:27than
32:27human
32:28beings
32:29and
32:29thirdly
32:30I
32:30predict
32:30it
32:30will
32:31be
32:32able
32:32to
32:33act
32:34and
32:34all of
32:35this
32:35if you
32:35take
32:35the
32:35consequence
32:36of
32:36that
32:36means
32:37that
32:37you
32:37will
32:37not
32:37have
32:38a
32:38co-pilot
32:39you will
32:40have
32:40a
32:40co-worker
32:41and
32:42and
32:42maybe
32:42half
32:43of
32:43your
32:43colleagues
32:44will
32:45be
32:46AI
32:46colleagues
32:47that
32:47are
32:47super
32:48intelligent
32:48everywhere
32:49and
32:49able
32:50to
32:50act
32:50and
32:51we
32:51need
32:52to
32:52redesign
32:52the
32:53way
32:53we
32:53think
32:54about
32:54leadership
32:54and
32:55organizations
32:56to take
32:57full
32:57advantage
32:58of
32:58this
32:58I
32:59come
32:59back
32:59it
33:00is
33:00the
33:00biggest
33:00opportunity
33:01we've
33:01ever
33:01had
33:02but
33:02one
33:03that
33:03needs
33:03clear
33:04and
33:04very
33:05conscious
33:05leadership
33:06very
33:07responsible
33:08leadership
33:08AI
33:08is
33:09not
33:09about
33:09stealing
33:10people's
33:10data
33:10and
33:11destroying
33:12democracies
33:13it is
33:14about
33:14making
33:15a better
33:15future
33:16for human
33:16beings
33:17terrific
33:19one
33:20advice
33:20the way
33:21we think
33:21about
33:22it
33:22at least
33:22the
33:23manpower
33:23group
33:23is
33:24we
33:24don't
33:24develop
33:25an
33:25AI
33:25strategy
33:26we
33:26develop
33:27a
33:27business
33:27strategy
33:28enabled
33:28by
33:28AI
33:29and
33:30then
33:30human
33:31centricity
33:32ethical
33:32AI
33:33is
33:34what's
33:34going
33:34to
33:34ultimately
33:35determine
33:35the
33:35success
33:36the
33:36fact
33:37that
33:37the
33:37technology
33:38is
33:38moving
33:38very
33:39very
33:39quickly
33:40is
33:40not
33:41a
33:41good
33:41predictor
33:42of
33:42what
33:42will
33:42happen
33:42what
33:43we
33:43as
33:43humans
33:43use
33:44it
33:44for
33:44is
33:45what's
33:46going
33:46to
33:46determine
33:47the
33:47future
33:47of
33:47work
33:48and
33:48if
33:49you
33:49keep
33:49those
33:49two
33:49things
33:50clear
33:50human
33:51first
33:51digital
33:52always
33:52and
33:53infuse
33:54it
33:54into
33:54your
33:54business
33:54strategy
33:55I
33:55think
33:55that's
33:56the
33:56fastest
33:56way
33:56to
33:57success
33:57and
33:57progress
33:57terrific
33:58Maude
33:59last
33:59word
33:59and
34:00keep
34:00that
34:01human
34:01touch
34:01which
34:02is
34:02very
34:02close
34:02from
34:02what
34:03you
34:03said
34:03because
34:03what
34:03we
34:04seen
34:04is
34:04sometimes
34:04we
34:05can
34:05have
34:05some
34:05efficiency
34:06gain
34:07what
34:07we
34:07seen
34:08is
34:08our
34:08clients
34:09want
34:09that
34:09human
34:09touch
34:10whether
34:10it's
34:10in
34:10a
34:11store
34:11because
34:11you
34:11know
34:11the
34:12person
34:12or
34:13it's
34:13because
34:13that
34:13handbag
34:14has
34:14been
34:14made
34:15hand
34:15to
34:15hand
34:15from
34:15the
34:16same
34:16person
34:16in
34:17the
34:17same
34:17atelier
34:17that
34:18they
34:18put
34:18value
34:19and
34:19that's
34:19how
34:20we
34:20create
34:20value
34:21as
34:21well
34:22Terrific
34:22Thank
34:23you
34:23all
34:23that
34:32Thank you
Commentaires