Pratique du Shadowing: How Disruptive Will Generative AI Be? A Micro-Level Analysis of Evidence and Expectations - Apprendre l'anglais à l'oral avec la vidéo

Choisissez un mode d'entraînement

Chaque mode ouvre une salle d'entraînement en plein écran.

315 phrases

Si les phrases sont trop courtes ou trop longues, cliquez sur Edit pour les ajuster.

Right, well, up until we sort out the screens and the presentation, the talk is about AI and disruption.

And before I start, I'd like to, as we sort out technology, I'd like you to think about the two keynotes that we have heard in this room.

Think about the keynote this morning, a keynote about sunniness and optimism and about the disruptive power of AI

that is about to sweep us from under the feet.

It's going to be deserving of its own nomenclature.

We'll live through it and it'll change us all entirely.

It'll democratize entrepreneurship and it's going to lead to a brighter future.

Think about the keynote that we heard yesterday.

90 % get it wrong.

10%, if we are generous, might get it right.

And we are kind of trying.

And if they have everything sorted out, if they have their strategy sorted out, we're here, then things will pan out.

That is a rather starkly different image from two industry people who either sponsor or are related to this event.

That tells you something.

And it tells you where we are, because there is a very interesting bimodal reality.

And I can tell you that when you speak with executives who speak to the press, you hear what you heard this morning.

When you hear with senior people behind closed doors, you hear a very different story.

And that's part of my motivation.

Where the heck are we?

Is AI going to disrupt us or not?

Why do we have these seemingly different structures about the world?

On the one hand, everything is changing.

On the other hand, there is very little that we can see on the aggregate.

There are a number of do -gooders.

I do apologize to MGI.

I don't quite apologize.

I think that they've lost um any sense of rigor um uh

that say we're all going to be rich rich rich rich rich

so there's this paper

that i saw um uh plant saying profits are going to go up by 4 .4 trillion dollars

which is rather bizarre you know i i asked actually actually

michael chew in a cc the guy running mgi said guys

i just these numbers don't make sense how did you get to them

and the answer is well we looked at productivity improvements we

multiplied it by a number of countries a number of companies uh the assets under management

and the entire world it's a big number i'm like okay

that sounds sweet but your strategy consultants this

if it is available to everyone would have been efficiency wages

that um uh that would have been competed down is is there anything

that are missing and the answer is not quite

and i understand it

because the people who are involved in the industry who want to tell you

that you should open your wallets by the way the allocation

of capital in terms of AI projects is not put in

under up until now at least the same category as any other capex

or opex business

and it does give a great justification to say the world

is changing we need to change as well on the other hand

if you start looking at things like the Gartner cycle you're going to see

that we are at the very peak of the hype cycle people get excited

when we get excited we get frothy

when we get frothy we lose our mind

and our role as academics is to ground the conversation bring some data

and then say what the heck is happening provided

that we do not have conflicts of interest

and we want to understand what's going on well i'm going

to give you already a way part of what the story is

and we're seeing that there are some doubts

that are raised even at the level of pilots do they pan out

or do they uh not pan out but um i think that the challenge,

and yesterday some of you may have heard my grumble with the O -Net focus,

the emphasis has been on looking at things that exist at the task or the modular or the professional level.

Now this is important and I'll tell you that's important

and it may have been the main reason that our field is so confused about the potential disruptive impact of AI.

We are looking at things at the level of the task

and then in the O -Net studies we aggregate tasks to profession

and we say whoo big change is gonna happen

because professions are changing or we look at an organization we look at the task and they think whoo

big change are gonna happen because what are organizations other than bundles of tasks

and I'm gonna take issue with

that now I have to say before thinking about the theoretical

has last year I was trying to figure out what exactly the disruptive impact of AI is

and I realized that we didn't have any systematic study

that looked not at whether AI is good at doing a particular task

and I'm not speaking about authors writing email on planes I'm speaking about tasks

that may be even more intricately involved in what organizations And there wasn't much.

So I said, OK, thankfully, the UK...

uh regional innovation fund funded a big study where i said look i'm going to go and i'm going to

explore whether directors in the uk see disruption as a result

of ai let's not look at the task let's look at organizations

and let's look at the people who run organizations

and then say hey is anything changing or not and i think

that this is important

because we need to start asking the questions at the level of the organizations

and of the things that the organizations produce rather than look at the tasks

and the things that the tasks magically aggregate because they don't.

Now, what did we look at?

Let me give you a bit of a sense of the descriptive statistics of at least

when we ask the question of is this working for you, what people answered.

By the way, this is ancient data.

It's a year old.

We did do some new data gathering to capture what's going on more recently.

and what was what we saw is

that out of the people who were using it about a

third didn't use it officially of course people used it

but the organization didn't formally use it

and about a third say we use it

and we don't see any impact about a third say yes we see it moving the needle

and somewhere in the middle and we're like great we have enough variance in order to see what

100

distinguishes those that get excited and those that do not get excited

101

if you look at descriptive of statistics what do you see well you see

102

that if you think about the maximum use cases of ai

103

and if you look at the extent to which conditional on that being used it is used systematically

104

rather than randomly you see some interesting patterns like those who aren't excited don't use it much

105

but also more importantly don't use it systematically and especially if you move from the moderate impact to the high impact,

106

it's much of a question of depth and consistency than a matter of breadth.

107

The other thing that we saw is that the closer you go to the core of the organization, are you not writing emails and doing all the things that everyone is looking at,

108

the more you see the impact being important when you go to operations,

109

when you go to IT and to the heart of the organization as opposed to the things that quite often IT,

110

sorry, AI is used for.

111

The other thing that we saw is that what matters is what you use it for.

112

Do you use it for cost reduction and productivity improvement?

113

Do you use it for hyper -personalization?

114

Or do you use it for using new revenue streams?

115

And it turns out that the more you use it for more strategic stuff, hyper -personalization, creation of new revenue streams, the more you see that this has an impact.

116

But that could only explain a few things in terms of the current usage.

117

What we were interested in is the understanding of the executives

118

in terms of their perception on whether AI would displace them or not, would break up the sectors or not.

119

So we started working at some detail.

120

In a moment, I'll show you the details of how we asked them and how the data was gathered.

121

and then we said well let's think about what's changing a couple of things

122

that came up i thought were interesting the first thing is

123

that we started saying what the heck do we mean by the word disruption

124

so when we think about disruption disruption means potentially three things the first thing

125

that it could mean is displacement are we all going to be dead people in marketing

126

and advertising are concerned rightly because they are some of the ones

127

that we have seen be the most exposed and the most liable to be displaced by Gen AI.

128

The other though was another two things.

129

The first one was disruption in a classic sense, in the sense that the criteria of what is good and what is not good are going to change in the sector.

130

And the third thing is the possibility of using this thing in order to be rich,

131

i .e there is a new way for us to differentiate ourselves.

132

And we started unbundling that and then we started correlating it with stuff

133

and then we saw as you'll see in a moment

134

that there are some features

135

that were more important than others now that's at the high level the what data we're looking at

136

but let me start jumping down to uh the research project

137

itself what did we do um what we did is that we started by figuring out what the heck is happening.

138

And as this is an area that requires us to get closer to the ground in order to articulate the hypothesis,

139

we used abduction and we had 96 people in six who are directors in six workshops,

140

an hour and a half to two hours, trying to listen so that they can tell us where they see disruption happening in their sector,

141

where they don't, so that we know what questions to ask

142

so that we know what are the variables to consider and so that we know what's going to be important.

143

then we had a survey

144

which went to the entire membership of the institute of directors in the uk

145

which was answered by 217 directors long survey about 40 minutes

146

we have similar data i worked with one of the major private equity firms uh

147

and the private equity results are similar we can't show them for legal reasons

148

but they are quite comparable to this

149

and then the third one is rather than do the usual academic stuff

150

which is to say I got my data I'll sit back

151

and I'll make sense of them we looked at the data

152

and we asked the people who had filled the survey to say why can you think

153

that we have these results so

154

that rather than us thinking what other mechanisms we ask people

155

what they thought was going on in order to be able to tease out more information about what is happening

156

so that we don't impose our own views on reality

157

but rather so that we can see how they can tell us what's going on here.

158

Why is this interesting and different?

159

So in terms of timing, we're starting a little bit late because I was whiffing for the first three, four minutes.

160

Thank you.

161

Excellent.

162

So the first thing, the first thing to say is that as I was saying, the body of work that we have, think about Eric Bringelson's work and the rest of it,

163

are looking at tasks at the modular level. And

164

if you think about the other things like the the big

165

BHI study again it looks at the production of slide decks for consultants

166

and the way that this evaluated which is interesting

167

but this is not what organizations do

168

and what we want to look at is at the level

169

of the value proposition of the entire organization not on any particular modular subtasks

170

that exist and what I would suggest is that like we have seen in previous iterations of technology,

171

when we're trying to understand the solo paradox of why IT did not have an impact, and people like Lauren Hitt and Eric B were saying, oh, it combines with other things.

172

Similarly, what we have is that Gen AI combines with other things, and I think, again, the cardinal mistake that we're making is to confuse the world for a modular,

173

a set of modular bundles which is not quite the case

174

so what we saw in the first set of roundtables is

175

that the important thing to ask is at the level of

176

the firm what is the value proposition of the business

177

and as such what are the enablers of the tasks

178

and the barriers that correspond to this value proposition and how

179

that changes either the firm or the sector then on the basis of

180

that we had the survey and the survey came up with some interesting facts.

181

Let me just give you very briefly some of the numbers that stood up.

182

We saw that some of the features that explained whether people were concerned or not about displacement were things like regulation,

183

modularity, and the existence of proprietary data.

184

So on the basis of this evidence, we said, let's try to figure out why.

185

We already had some hypothesis, but then we said, let's create these roundtables,

186

and we had 160 -odd people who came in these nine

187

roundtables where we bundled people in terms of people who had high -propriety data or low -propriety data,

188

again, different sectors trying to help us understand why the data in these sectors were of a particular form.

189

And here's what we found.

190

We found that people were speaking much more in terms of how Gen AI can complement and displace the value proposition overall.

191

What do I mean about that?

192

Well, I mean that there are some things that make the value proposition defensible or not.

193

If you start looking at the differences and you look at people saying, well, I'm afraid that we're going to be displaced versus, no, I'm not terribly afraid that I'm going to be displaced.

194

So to give you a more concrete example, if you think about different law firms, there's one law firm that says, well,

195

what we think we're doing is that we're offering a set of bespoke solutions

196

that requires us to understand what the customer really wants.

197

And because this is not really displaceable in terms of gen AI, we don't think we're going to be unbundled.

198

Or another firm said, we think that the legal liability

199

that exists is a very potent protector so that we can essentially use AI, but it's going to be within our own limit.

200

And that is what's not really driving it.

201

That's why people in accounting are more comfortable than people in

202

consulting in terms of their ability to withstand the potential disruptive shock when it comes to Gen AI.

203

We then did a number of things, kicking the tires, and we said, well, what if we only looked at the snapshot of history?

204

So that's why we had another big workshop in June of this year,

205

and we reviewed, we had a number of further detailed surveys to say,

206

have things changed with the emergence of chat GPT -4?

207

The general answer is no. And it's interesting because especially in the qualitative surveys,

208

the features that existed continue persisting and suggesting that there is rhyme and reason in that regard.

209

The other thing that we looked at is we were speaking, we were looking at those who felt that they are more differentiable

210

because they are doing things that are more unique and more bespoke.

211

And of course, we said, is it the case that these people could simply be full of themselves and they think that they're awesome,

212

but they are just giving us their own perception that they are invincible?

213

So we went back, we tracked everyone via LinkedIn, we then used the corporate websites, we used the Wayback Machine,

214

and we saw, do the people who think

215

that they are more bespoke have websites and information that is clearly different from the rest of the industry, so there is a prima facie validity,

216

so that we have some external validity?

217

And the answer is yes.

218

So what is it that we have learned from that?

219

What we've learned from that is that when you think about whether Gen AI is going to disrupt you or not,

220

you need to be thinking about whether it'll disrupt the value proposition.

221

Looking only at whether it'll affect an individual activity

222

that exists within your own boundaries is not something

223

that is going to tell you whether disruption will happen or not

224

so we need to shift from thinking about what happens within

225

the modules to thinking about where the modules are situated on

226

so that the value proposition is important the gen ai based disruption is an interesting combination about of both displacement

227

and new differentiability and we see this as the challenge in the sector

228

but also the opportunities for others we do think that

229

that will lead to a serious reordering in a number of sectors

230

so there will be a shock ironically not a shock of everyone identically

231

because technology is eating humans a shock to some firms

232

that are more vulnerable and others who are going to take their position

233

so ironically yes there will be a disruption but of a different sort of a strategic short that's going to be

234

reorganizing the sector modularity emerges as a very important feature and

235

that might be why people like ethan are so excited about gen ai

236

and the people who are looking at the macro can't see the results of the data

237

because the world does not consist of modular sectors

238

but some are modular and decomposable so these are going to be less defensible

239

and we can start having a more sophisticated analysis of who's going to be threatened

240

and who not and the role of regulation is also super important

241

because it also defines the things that are potentially going to be challenge or not,

242

which is really interesting, because things that would have looked as boring and

243

standard as the rules of regulation may end up being extremely important in a world of Gen AI.

244

That's the very quick overview, and I look forward to your questions.

245

Thank you very much.

246

Yes?

247

Yes?

248

It seems like, to me, like, what's current, is patent -based competition?

249

Well, I mean, look, the reason that I think that you should look at organizations, and whether organizations, organizations overall, business units, not necessarily corporates.

250

By the way, the corporate is another very interesting story.

251

And there's a parallel project that we're starting now, which is what is a cohesion in a world of Gen AI?

252

And how can you think about the value of multi -business firms when resource allocation can be supported and supplanted by that?

253

So it's another interesting question.

254

But again, you're starting to look at essentially the units that exist that produce P &L results.

255

And then you say, are there, is there rhyme and reason in the way these units are organized as opposed to what's happening in the task?

256

I'm not suggesting that tasks are irrelevant.

257

Of course, they're going to happen, right?

258

We have the same thing when we're looking at innovation.

259

I was a student here at Wharton of Sid Winter and Sid was saying, don't look only at.

260

issues that can be studied when you look at patent dynamics, because there's a number of other sectors that cannot have intellectual protection that is captured by that.

261

And then we grew up and we started looking at innovations, not only looking at citation counts in areas where patents are important, saying, let's think about innovation sort of more broadly.

262

I hope that a similar thing will happen in the world

263

that we're looking at by considering essentially company by company or or value prop by value prop, and how this is being affected by Gen .ai.

264

So my encouragement is we have already a lot of the things that are easy and that generate PhDs or generate publications, which is I'm going to take tasks,

265

I will take Onet as my guide, and then hopefully it will all pan out, and there are going to be reviewers and editors who have done similar things and it's going to be published.

266

I get it.

267

I'm an editor too.

268

But on the other hand, what I'm suggesting is that if we are to be useful as a field, we need to be a little bit more courageous and we need to say,

269

let's look at where the action appears to be at organizations

270

and look at it at the value prop and how AI affects the value proposition.

271

I think that we will come up with a more nuanced version.

272

We will be able to link a little bit the two keynotes saying the first is cool, but cute and applies only to AWS one,

273

only to a small part of the world

274

and a big part of the world is facing the problems that the lady from Accenture was speaking to us about, right?

275

So I think that this is my hope

276

that we're going to start complementing the work

277

that happens at the task level with a lot with the work

278

that happens at the value proposition or the business unit level but that's just my wishful thinking.

279

Will Will we do it?

280

I don't know.

281

It's going to be hard.

282

So here we have a mixed blessing,

283

because the people who wanted to spend two plus hours to come to a meeting at LBS

284

or spend 40 minutes answering the survey are a non -representative sample of those who are worried about it

285

and whose motivation is to figure out what other people want to do.

286

If you think about what we're looking at is a part of the population that is more at risk of being excited.

287

And if you look at the entire world, this is a small part of the overall population of people, many of whom are going to say,

288

all of that is hype and I don't care about it.

289

So I can't speak on what is happening for those who are simply closing their eyes and their ears saying, I think that this is just hype.

290

I don't want to look at it.

291

I am not literate.

292

So I think that there is going to be an interesting challenge.

293

I think that what's going to happen is that when you see organizations that are able to benefit from it,

294

perhaps either through national selection or through some learning, they might adjust.

295

But I think that you're going to see a bimodal or multimodal set of organizations.

296

I expect some AI -powered firms to compete with some firms that are able to integrate the AI well, and others who are hobbling along,

297

and the relative shares of these three groups are going to change over time.

298

true but what i will leave you with is the implication of our study isn't

299

that everyone should invest a lot and immediately on AI.

300

It is that AI is going to be a threat for some people, a lesser threat for others.

301

Another project that we're starting now with the Henson Institute and BCG is to look at the different types of knowledge work.

302

And what we're already finding is that there are different species in terms of those who are well or poorly adapted.

303

It's acceptation, it's a pre -adaptation.

304

There are some that essentially will end up being well by accident, by coincidence, well adapted to integrate AI and others that will not, right?

305

So there's going to be an interesting variation, interesting systematic variation.

306

And also I would say it isn't the fact that everyone needs to rush to entirely stop everything,

307

uproot the things that they're doing and just rethink everything on the base of AI.

308

This would not be the right advice.

309

that we as academics would give.

310

Venters, yes, sure, that's what they need to tell you.

311

But we as academics should provide a more balanced answer saying,

312

let me tell you how you look like and as such what on the margin is going to be better for you.

313

And this is where I hope that academic research is going to shed more light.

314

That's just the first step and much more to follow, probably in meetings such as these.

315

Thank you very much.

Télécharger l'application

Tout pour parler couramment

Prononciation IAChaque phrase notée

Phonétique (IPA)Maîtrisez chaque son

VocabulaireEnrichissez votre lexique

Jeu de vocabulaireApprenez en jouant

Scanner le QR

Download on theApp Store Get it onGoogle Play

FOR YOU

Vidéo recommandée

74 phrases

Vocabulaire et conseils d’expression pour cette leçon

Cette leçon d’expression orale s’appuie sur la vidéo « How Disruptive Will Generative AI Be? A Micro-Level Analysis of Evidence and Expectations ». Cette vidéo contient 315 phrases et 4237 mots à répéter en shadowing. La partie parlée dure 25:23. Le locuteur parle à un rythme naturel d’environ 167 mots par minute, proche d’une conversation courante. 85 % des mots font partie des 3 000 mots les plus courants en anglais ; le reste mérite d’être vu avant de commencer.

Vocabulaire clé de cette vidéo

15 mots de la vidéo à retenir, avec leur prononciation et leur sens :

Mot	Prononciation	Sens
disruption nom	/dɪsˈɹʌpʃən/	interruption, perturbation
proposition nom	/ˌpɹɑpəˈzɪʃən/	proposition
modular adjectif	/ˈmɑd͡ʒəlɚ/	modulaire
shock nom	/ʃɔk/	choc
essentially adverbe	/ɪˈsɛnʃəli/	essentiellement, en essence
innovation nom	/ˌɪn.əˈveɪ.ʃən/	innovation
regulation nom	/ˌɹɛɡjʊˈleɪʃən/	règlement, réglementation
patent nom	/ˈpeɪtənt/	brevet
hype nom	/haɪp/	matraquage publicitaire, promotion
aggregate nom	/ˈæɡ.ɹɪ.ɡət/	agrégat
prop nom	/pɹɑp/	cale, support
disrupt verbe	/dɪsˈɹʌpt/	perturber, déranger
strategic adjectif	/stɹəˈti.dʒɪk/	stratégique
capture verbe	/ˈkæp.(t)ʃɚ/	saisir, capturer
potentially adverbe	/pəˈtɛnʃ(ə)li/	potentiellement

Les verbes à particule que vous entendrez

Mot	Sens
figure out verbe	cerner
sort out verbe	résoudre
come up with verbe	inventer, imaginer
end up verbe	se terminer, se finir
break up verbe	rompre, couper
come up verbe	se montrer, sortir du bois
go back verbe	rentrer, retourner
go up verbe	monter, s'élever

Des phrases à répéter

Des phrases courtes et complètes de la vidéo, à réutiliser dans la conversation de tous les jours :

Where the heck are we?
Why do we have these seemingly different structures about the world?
Do you use it for hyper -personalization?
What do I mean about that?
It seems like, to me, like, what's current, is patent -based competition?

La grammaire de cette vidéo

Les structures que le locuteur utilise le plus, avec les mots exacts de la vidéo :

Structure	Dans la vidéo
Voix passive be + participe passé — l’accent est mis sur ce qui arrive, pas sur qui le fait	have been competed · are involved · is not put
Propositions relatives who / which + proposition — une précision sur une personne ou une chose	people who either · people who are · people who run
Present perfect have/has + participe passé — une action passée qui compte encore maintenant	have heard · they've lost · has been

Prononciation à surveiller

Le locuteur utilise 79 contractions et formes réduites, comme we're, I'm, don't. Prononcez-les sous leur forme courte, telles que vous les entendez.

Les sons « th »: gather /ˈɡæðɚ/, hypothesis /haɪˈpɒθɪsɪs/, theoretical /ˌθi.əˈɹɛt.ɪ.kəl/, threaten /ˈθɹɛt.n̩/, withstand /wɪðˈstænd/
Les sons « sh » et « zh »: disruption /dɪsˈɹʌpʃən/, proposition /ˌpɹɑpəˈzɪʃən/, shock /ʃɔk/, essentially /ɪˈsɛnʃəli/, innovation /ˌɪn.əˈveɪ.ʃən/
Mots longs — placez bien l’accent: proposition /ˌpɹɑpəˈzɪʃən/, essentially /ɪˈsɛnʃəli/, innovation /ˌɪn.əˈveɪ.ʃən/, regulation /ˌɹɛɡjʊˈleɪʃən/, potentially /pəˈtɛnʃ(ə)li/

Les sons difficiles pour les francophones :

/h/ — il se souffle, il n’est pas muet: heck /hɛk/, hype /haɪp/, hypothesis /haɪˈpɒθɪsɪs/
/tʃ/ et /dʒ/ — à ne pas adoucir en « ch » et « j »: modular /ˈmɑd͡ʒəlɚ/, strategic /stɹəˈti.dʒɪk/, generate /ˈd͡ʒɛn.ə.ɹeɪt/, apologize /əˈpɑləˌd͡ʒaɪz/, module /ˈmɑd͡ʒul/
/r/ anglais — langue recourbée, sans frotter la gorge: disruption /dɪsˈɹʌpʃən/, proposition /ˌpɹɑpəˈzɪʃən/, regulation /ˌɹɛɡjʊˈleɪʃən/, aggregate /ˈæɡ.ɹɪ.ɡət/, prop /pɹɑp/

Comment s’entraîner avec cette vidéo

Écoutez la vidéo en entier une fois sans parler et notez les mots que vous ne connaissez pas.
Commencez à la vitesse 0,75×, répétez phrase par phrase, puis revenez à la vitesse normale quand cela devient facile.
Enregistrez-vous et comparez avec l’original, en faisant attention à des mots comme disruption, proposition, modular.

Qu'est-ce que la technique du Shadowing ?

Le Shadowing est une technique d'apprentissage des langues fondée sur la science, développée à l'origine pour la formation des interprètes professionnels. Le principe est simple mais puissant : vous écoutez de l'anglais natif et le répétez immédiatement à voix haute — comme une ombre suivant le locuteur avec un décalage de 1 à 2 secondes. Les recherches montrent une amélioration significative de la précision de la prononciation, de l'intonation, du rythme, des liaisons, de la compréhension orale et de la fluidité.

Technique du shadowing : lire le guide complet étape par étape →

Pratique du Shadowing: How Disruptive Will Generative AI Be? A Micro-Level Analysis of Evidence and Expectations - Apprendre l'anglais à l'oral avec la vidéo

Choisissez un mode d'entraînement

Télécharger l'application

AI & Inequality | Daron Acemoglu: Big Tech poses risks as AI reshapes society

The Only SaaS Moats Left in 2026

Harnessing generative AI in architecture, engineering & construction

No, Really — Is A.I. a Bubble? | The Ezra Klein Show

Inside the secret race to build artificial intelligence | The Gray Area

Elon Musk: The End of Work, the Death of Money, and a World Run by AI

How AI is helping brands identify consumer trends faster

Is ML Engineering Dead? (AI vs ML Engineer)

AI Governance Explained: Ethical, Transparent & Accountable AI for the Future

Are we heading to an era of stagnation or innovation?

China just dropped a new frontier model... and nobody can explain how

Alexandr Wang on what matters in an age of abundant intelligence