- il y a 20 heures
Tech and Sovereignty, A Strategic Dilemma
Catégorie
🤖
TechnologieTranscription
00:00Bonjour tout le monde, je suis un partenaire global avec EY et un peu de choses que je veux ajouter
00:06c'est que je suis un canadien, donc je suis très heureux d'être ici pour le Canada.
00:10Et deux, j'ai passé un grand nombre de mes carrières avec des technologies et des entreprises
00:15donc c'est un vrai privilège de parler avec le CEO de l'industrie, Christelle Orange.
00:23Merci.
00:23Merci.
00:25Sovereignty is becoming a big part of the narrative, the talk track.
00:30When it comes to AI, when it comes to technology, when it comes to innovation,
00:33sovereignty is taking the forefront of the equation.
00:36A few months ago we were talking about data, we were talking about energy.
00:40Now with the emergence of AI, we're talking as well about the compute element,
00:46we're talking about the culture, the jurisdiction, the language element of sovereignty.
00:53So why is there so much talk about tech and sovereignty and why is it often called dilemma?
00:59Well I think, let's be realistic, if we look at the digital value chain, there's a lot of dependencies.
01:07And if we look at it from the European point of view actually, the strategic autonomy of Europe,
01:12when we think about digital, is actually very poor.
01:15And we know that.
01:16You just look at your mobile phone.
01:18I mean the OS, the operating system of your mobile phone actually, there's two options.
01:24Two US companies.
01:26You look, I can go down to the chipsets of course.
01:29We can talk about of course the, I mean the cloud.
01:32And we know the three big hyperscalers or Chinese options.
01:36So there's a lot of dependencies.
01:39But I don't think when we think about sovereignty, what we call sovereignty at Orange is not about,
01:47I mean, closing the borders and thinking about disconnecting, because that's impossible in the digital world.
01:55The way we look at it, we've defined it with four C's.
01:59It's about control, it's about choice, it's about competencies, and it's about critical size.
02:07So control is about securing infrastructure, making them resilient.
02:13As a telecom operator, we know our services are so essential.
02:18Securing, making our services resilient, making our infrastructure resilient.
02:22Whether we're talking about the fantastic events like the Olympics or VivaTech,
02:28or we're talking about disastrous recovery when there's flooding, when there's storms.
02:34So securing infrastructure, and that's about controlling your infrastructure.
02:39Choice, it's about reversibility, having options, ideally having European options in some cases,
02:47and that's why we do a lot of things, and VivaTech is a great example of,
02:51and we can talk about AI as well.
02:54And then competencies, competencies is very critical, because the first thing when we talk to our customers,
03:01or when we talk to administrations, governments, if you don't have the skills and the know-how to understand,
03:08first of all, where you have gaps, what you can do yourself.
03:11We're investing on open source, but to invest on open source, you need to have the competencies, the skills, the
03:17know-how.
03:17So that's the fourth one about competencies.
03:21And then probably the most important in Europe, critical size.
03:25If we have tech giants coming from China or America, that's because they have a home market that allows to
03:32scale.
03:33The issue in Europe is that we are so fragmented.
03:36So as Orange, of course, we are trying to be the number one in every market where we operate.
03:41We're number one in France, and we want to stay number one.
03:44We're now number one in Spain.
03:45And we are consolidating, and I think we need to further consolidate the telecom market,
03:50because there's no magic.
03:52In the world of software, scale matters.
03:55And so that's about this fourth T about critical size.
03:59That's great.
04:01It's obviously we need to find the right balance between the agility that's needed and just driving decisions around AI
04:10and making sure that the decisions are not taking too long to the point where they're slowing down our competitiveness
04:20in general.
04:20And the way I see it is that technology is the backbone of industry,
04:26connectivity is the backbone of technology, and telcos are the backbone of connectivity as well.
04:32Any major lessons learned when it comes to sovereignty that Orange or the telco sector could learn from?
04:40Well, there's many things we're doing.
04:43The number one solution, I mean, the number one, I would say, strategic decision,
04:46we want to master and own our infrastructures.
04:51And infrastructure means we own submarine cables.
04:54We own submarine, I mean, not submarine, but vessels to repair submarine cables.
04:59We own, of course, I mean, networks.
05:03And when I say it doesn't mean owning on your balance sheet, even though we own a lot,
05:08it's also having the know-how again to repair.
05:10So it's about mastering your value chain.
05:13We're investing in some areas.
05:15Take cloud as a good example.
05:16Of course, we work with hyperscalers.
05:19I mean, we're a global company.
05:21We want to use the best technology to move fast to innovate.
05:24So we work with hyperscalers.
05:26But when it comes to some very critical data that we own,
05:30we have private cloud infrastructure.
05:32When we think about the next generation of our network, that software hosted on cloud,
05:37we are investing on open source technology.
05:40So we are building our own telco cloud.
05:42We just don't do it for Orange.
05:44This is something we're working on with all the telecom operators in Europe.
05:48Now, if we are talking about trusted clouds,
05:52and this is something where we are investing together with Capgemini on blue in France.
05:57So that's bringing the best of technology, in that case from the Microsoft Azure Stacks,
06:02in an environment that's trusted for companies that have to meet the second cloud needs.
06:08Because in some businesses, and telecom are highly regulated,
06:12we have to meet with the local regulation.
06:15And we do that country by country.
06:17So, of course, we tend to think here we are in Paris.
06:20So, of course, we talk a lot about French tech and French sovereignty.
06:23The reality is that the more we connect our networks, the more resilient they are.
06:28So I think thinking about closing borders is not the right way.
06:31That's not the way we do it.
06:32We are adding connectivity to our country networks because that's extremely resilient.
06:37If you think about what happened in Spain a few weeks back, blackout for hours,
06:43you realize how critical power is, that we know, and how critical connectivity is.
06:49And that's something that we will continue to work on.
06:51Now, of course, geopolitics is now everywhere in every board discussion.
06:56And that's true.
06:58And we work with Chinese partners.
07:01We work with American partners.
07:03When I meet with country leaders in Africa,
07:05I guarantee you that in Ivory Coast or in Senegal,
07:09country leaders, they don't want to rely on Europe.
07:12They also want to invest in startups, in their ecosystem.
07:16They want to build the know-how.
07:17And that's something we're doing as well.
07:19That's great.
07:20Shifting gears very quickly to AI.
07:23I can't believe it took us seven minutes before we got into AI.
07:27Scaling AI is still not too simple.
07:30It's based on fragmented data.
07:32It has legacy systems.
07:34The workforce is not completely up to speed.
07:37And the regulatory environment is not easy to figure out yet.
07:41And very fragmented.
07:43How is Orange?
07:44Or can you share your perspective on how Orange is leveraging trusted AI model
07:48on a few fronts?
07:50Obviously to support digital sovereignty,
07:52but also to enhance customer experience and to improve overall operations.
07:58So AI, and I think last year I was on stage to talk about how fast AI adoption was among
08:04consumers.
08:05And that puts a lot of pressure on any large organization because our employees are consumers.
08:10And so when young people join a new company, they've been using ChatGPT or LeSha or Gemini, whatever.
08:20They come to a company, they expect to have the same great experience that they have in their personal life.
08:25So of course we're adopting solutions internally for our employees.
08:28So we've developed a secure environment.
08:30We've developed a solution.
08:32We call it live intelligence where our employees can safely use the company's data, the company's information,
08:40any PowerPoint document, you take it, and any type of document that's internal document,
08:45and then use it, choose which language they want to use.
08:49And, of course, we're working with Mistral.
08:52We're very proud to have them.
08:54But we're also working with other U.S. providers.
08:57And that's about using the technology because the worst thing would be to ignore
09:03and not to get used and not to learn from those use cases.
09:06Now, of course, when it comes to our customer experience, our call centers, our marketing, our networks, our cybersecurity,
09:13here we have programs at company level to try use cases, to leverage them, and to scale them.
09:19That's not easy because the reality is that a lot of the data, that's a discussion we've had for a
09:25decade
09:25when we talk about digital transformation, but the data sits for us in every country
09:30because by regulation there's a number of data that we have in our networks that we're not allowed.
09:37And that's for good reason because this is GDPR.
09:40This is about privacy and protecting our customers.
09:43But, of course, if we want to build the best model to monitor our networks across Europe,
09:47we need to learn from the network log and from all the accidents.
09:52So we have a global network operation center that's connecting all our networks across Europe,
09:57that's building this new algorithm, and that's making sure that we're building the new software stack
10:02that will help us drive our networks out that are immensely complex
10:07because we're just carrying the legacy of 2G and copper and then we are adding cloud and AI and fiber
10:15and 5G and private networks
10:17and now AI is everywhere.
10:19Cybersecurity is another example where without AI it's today just impossible.
10:24AI is used by the attackers.
10:27You need to use AI to defend, so to defend our customers, to defend and to enrich our portfolio of
10:33solutions for our Orange Cyber Defense.
10:35It does seem that many of the large organizations have built their own LLMs.
10:39We have what we refer to as EYQ and our 400,000 people have access to asking it questions as
10:46well.
10:46And the adoption has been surprising. Close to 85% of our employees have adopted it.
10:51A couple of angles maybe that could just touch on.
10:55So we have made the decision and we have started investing $1.4 billion in AI in general.
11:02And we're looking at it through three fronts.
11:03We're looking at it through the angle of us being client zero.
11:07So our employees need to be completely up to speed.
11:10And I'm assuming you're doing the same.
11:12But at the same time, our processes need to be completely revisited through AI.
11:17And the one thing I heard today is that an agent is a very young,
11:23is the equivalent of a very young intern with a lot of potential.
11:27And it's on us to figure out how we move that agent who's an intern with close to zero experience
11:34to a more seasoned executive or employee who can make the right decision.
11:38So there's a lot of eyes on making sure that the agents remain trusted
11:45and everything gets done the right way.
11:47And obviously, the second element is working with regulators to make sure
11:50that we shape the regulatory environment where we can.
11:53And the third angle is working with clients.
11:56Because to your point, clients have moved from last year saying,
12:00well, we're going to adopt AI through these two, three, four use cases,
12:03to now AI being embedded on multiple, multiple fronts.
12:08And the one thing I learned yesterday, by the way,
12:10the realization yesterday is that data centers are being taken over by gigafactories
12:20and by compute centers.
12:22Well, yes and no.
12:23Okay, perfect.
12:24Let me be provocative here.
12:26Because of course, everyone is focusing on the big gigafactory.
12:29And yes, training models requires a lot of compute.
12:34But let me tell you, our job when we manage networks is to get the right technology
12:39for the right use case.
12:40And we are making sure that we want to talk about frugal AI.
12:44Because it's a fight every day and night for our teams to keep the energy consumption when we have traffic
12:51increasing by 20% every year.
12:54And so our challenge is to make sure that we use, of course, we need the large models and the
12:58best of AI.
12:59And here, there's amazing innovation here when AI can now touch and see and sense and smell.
13:06And if you think about glasses and everything, that's going to require really smart networks.
13:11But I tell you, if everyone here gets smart glasses and is streaming what we see around to a big
13:19data center, that's not going to work.
13:22You need edge compute at the back stage of this room if you want this to happen.
13:27By the way, that's what we've done during the Olympics.
13:29You know, we've connected the boats with all the athletes on the river.
13:34And we've done that thanks to mobile networks.
13:37So we used, actually, private networks to do that.
13:40So, actually, the art of using the right model and getting compute at the right place.
13:46Now, there's no doubt we will need a lot more compute capacity because it's true that we keep on creating
13:52data at a completely unprecedented pace.
13:56And this keeps accelerating.
13:58And you all have your own experience.
14:00And you all have all your mobile phones full of pictures and full of video.
14:05And even though you take the same picture as someone sitting next to you, you still take two pictures on
14:10two mobile phones that go into the cloud.
14:12So we keep on with this exponential growth of data.
14:16But the challenge and the key question for us is the networks of tomorrow will embed compute.
14:22And it will be compute that will be at the right place and using AI for the right use case.
14:28And this is what we call frugal AI, which doesn't mean we don't want to use AI, but it's the
14:33right AI for the right use case.
14:36It's responsible AI as well.
14:38So we have time for one more question.
14:39And I have to ask, you obviously have a front seat on a lot of the things that are happening,
14:46whether the geopolitical tensions, whether the different parts of the world and how they're reacting to what's going on.
15:06And I think if you go to the orange booth, we talk about trusted networks, trusted experience, trusted customers, trusted
15:14everything because I think trust, this is something that's difficult to earn.
15:19And this is something that you can lose easily.
15:22And I don't think in this world, we all do business based on relationship.
15:26Trust is at the heart of everything.
15:28That's true in our family.
15:30That's true with our relationships.
15:31That's true in business.
15:32And I think in this world of technology, the challenge of AI would be what can you trust?
15:38What can you not trust?
15:39And this is something that we realize our life is now exposed in this digital world.
15:45Going backward is very difficult.
15:47So we have to protect children.
15:49We have to protect our private life.
15:51But building this trusted environment and providing an environment where, yes, our customers need to trust us.
15:59And we need to provide them also the keys to building themselves this trusted environment.
16:04That's really what we call responsible technology.
16:07That's part of our responsibility.
16:08And the advice to the next round of leaders is just make sure you're trusted, you do the right thing.
16:14Well, I mean, I want to be humble because I think we live in a very uncertain environment.
16:20And I think people look at us and they want a direction and they look at us and they say
16:24this is right, this is wrong.
16:26I think everyone needs to find its own way.
16:29And, you know, I'm French.
16:31I grew up in France.
16:32Yes, I've worked in international environment.
16:34But I have the biases of being born in this country.
16:37There's nothing good or bad about it.
16:40But when I talk to people from China, the US, Africa, I think being curious is probably the most important
16:47thing because social networks try to force us in our own bubbles.
16:52And I think we need to make sure we break barriers, we break the bubbles, we continue with human relationships.
16:59I think more and more, you know, some people pay to get a digital detox.
17:04Yeah.
17:04This is amazing.
17:06Turn off your phone, talk to your neighbor, go in the forest.
17:12I mean, this is also part of what I think the young generation and especially we know the challenges of
17:18our children.
17:18And we talk a lot about the impact of social networks.
17:21And I think this is something that we're building a next generation that grow up with digital technology, with mobile
17:30phones, with social networks.
17:31We need to help them.
17:33They are helping us as well, cope with that.
17:35But I think there's a lot of things happening in the society.
17:39But making sure we are curious and we're also conscious about our biases because we're all biased.
17:45So be curious, be conscious, be human.
17:49And the one thing you didn't say but it came out loud and clear, be humble.
17:52So thank you.
17:53Be humble, yes.
17:54Thank you.
17:55Thank you.
17:56Great conversation and I think our time is up.
17:59Thank you.
18:00Thank you.
18:01Thank you.
Commentaires