00:00Good afternoon everyone.
00:02Let me start with something very concrete.
00:06A few days ago, access to Fable 5, the super powerful LLM from Entropic,
00:14derived from the highly publicized mythos,
00:17was suspended for non-US people and companies in the name of US national security.
00:25I think we should say it clearly, this is not just another episode in the fast-moving AI story.
00:33It is a wake-up call.
00:36Because overnight, what many people still viewed as an open technological race
00:41suddenly looked very different.
00:44It reminded us that in AI, access is not always guaranteed.
00:49It can be restricted.
00:51It can be conditioned.
00:53It can be decided elsewhere, especially outside Europe.
00:57And when that happens, the question is no longer who has the best model, whatever it means.
01:05The question becomes, who remains in control?
01:10For Europe, for governments, for companies, this changes the conversation.
01:16It forces all of us to revisit some of our assumptions, and frankly, some of our priorities.
01:23For years, the debate around AI has been dominated by performance, bigger models, faster adoption, more use cases, more productivity.
01:35Of course, these things matter.
01:39But what the mythos episode reveals is that performance alone is not a strategy.
01:46If access to a critical model can be suspended from outside Europe, then AI is not only a matter of
01:52innovation,
01:53it's also a matter of geopolitics, strategic autonomy, resilience, and ultimately, trust.
02:00It means that both states and businesses need to ask harder questions.
02:06Not only how fast can we deploy AI, but also on what foundations, under whose rules, with what guarantees.
02:17Because a company that builds its transformation on tools it cannot fully access, explain, govern, secure, or economically control is
02:27not accelerating safely.
02:28It's increasing its dependency.
02:31And dependency in the edge of AI is a strategic risk.
02:36And because the economic and power stakes are so high, that promises alone have little values.
02:45So today, I'd like to focus on one simple question.
02:49How do we regain control without turning our backs to progress?
02:54I do not believe the answer is to slow down or to step back from AI.
03:01Quite the opposite.
03:02The answer is to build AI on foundations we can trust.
03:07First, we need to stop thinking about AI as something abstract, almost floating around us.
03:15AI is physical.
03:16It runs on data centers, network, cloud, environments, interconnections, energy systems, cybersecurity layers, talent, and operating models.
03:27An AI agent used by thousands of employees does not only rely on a model.
03:34It relies on secure connectivity, cloud capacity, identity management, data governance, and constant monitoring.
03:42If one of these layers fails, the service may still appear available.
03:49But trust is already broken and the consequences can be dramatic.
03:54And it also runs within political, legal, and economic boundaries.
03:59The Mythos decision made that visible in the clearest possible way.
04:05Second, regaining control means treating infrastructure as strategic.
04:11At Orange, this is not only theory.
04:15It's our daily reality.
04:17We operate at scale across 26 countries and serve 340 million customers worldwide daily.
04:24We rely on fixed and mobile networks, terrestrial and submarine infrastructure,
04:30satellite capacities, secure cloud environments, and cybersecurity expertise
04:34to keep critical digital systems running safely every second of every day.
04:41And this matters for AI.
04:43Because there's no trustworthy AI without trusted infrastructure.
04:50Third, controlling the model itself is not enough.
04:55You also need to control the access rights.
04:58And that changes everything.
04:59What matters is not only whether a model is powerful.
05:04What matters is whether you can access it when you need it.
05:09Whether you can deploy it under predictable conditions.
05:12Whether you can rely on it over time.
05:15Whether you have a deep understanding of its strengths and weaknesses.
05:19And of its potential biases.
05:21And strategic autonomy is also an economic issue.
05:26AI is increasingly consumed as a service, request by request, token by token.
05:33Through models whose economics are not yet fully stabilized for many businesses.
05:40Many companies have experienced this very concretely.
05:44A use case starts small.
05:47Adoption grows fast.
05:49And suddenly the bill grows faster than the value created.
05:54Not to mention the societal and environmental responsibility of such losses of control.
06:03If you do not understand what drives cost of AI.
06:07If you cannot connect usage to value.
06:10If you cannot anticipate the economics of your dependency.
06:14Then you are not fully in control of your transformation.
06:18And finally, there is another dimension where this loss of control becomes even more serious.
06:25Cyber security.
06:27Because AI is not only changing productivity.
06:31It's changing the nature of cyber risk.
06:35The challenge is not only that attacks may become more automated and powerful.
06:40The deeper challenge is that advanced models can identify chains of vulnerabilities, weak points, and systemic flaws at a scale
06:51and speed that human teams alone will struggle to match.
06:57That changes the rhythm of defense.
07:00And when the pace of discovery accelerates, the ability to understand, prioritize, remediate, and recover becomes even more critical.
07:10This is why for us, cyber security is not a separate conversation from AI.
07:16It's one of the places where the truth of AI will be tested.
07:22If you cannot understand your own exposure.
07:25If you cannot investigate your own systems.
07:27If you cannot secure your critical environments without depending entirely on external black boxes.
07:34Then you're not in control.
07:36And without control, there is no trust.
07:40Which brings me to my strongest conviction.
07:44The more AI advances, the more strategic the human becomes.
07:50We often hear that AI will replace people.
07:54I believe the opposite is true.
07:58The more machines produce, the more value shifts toward what they will never do.
08:06Judgment, nuance, the courage to decide under uncertainty, the empathy, the intuition of a network engineer facing a complex failure.
08:18The responsibility of a leader.
08:21A company that can no longer write, analyze, or decide without AI is not an augmented company.
08:29It is a vulnerable one.
08:31The next frontier of AI is not only technical.
08:36It is human.
08:38Not because machines are not powerful enough.
08:41But because I am firmly convinced that values, judgment, courage, and responsibility are not features that we can just download.
08:51Staying in control does not mean rejecting AI.
08:54Why?
08:55It means refusing dependency and ease.
08:59And that intellectual dependence, independence, or perhaps we should say that independence of judgment rests on physical independence.
09:10Control of your data.
09:12Control of your networks.
09:14Control of your critical infrastructure and your compute capacity.
09:17But above all, it rests on control of the models themselves and of the rights attached to them.
09:25And for Europe, I believe this is now the real frontier.
09:29Not simply building more AI, but building AI that we can access, govern, secure, and challenge on our own.
09:38Moving together faster and with direction is now required, while remaining open to the rest of the world, but without
09:46naivety.
09:48At Orange, that's how we see our role.
09:50We believe the next phase of AI will be defined by the ability to combine technological power, strategic autonomy, cybersecurity,
10:00and human judgment.
10:02We believe we have the opportunity to think long-term and not always overreacting to external challenges.
10:09That's how we will create lasting trust.
10:13That's how we will create resilience for us, for our partners, for our clients, and for our customers.
10:19And that's how we will keep AI a tool for open progress.
10:24Not a technological weapon to impose us a political and economic supremacy.
10:30So yes, the mythos episode is a wake-up call.
10:35But maybe it's also a very useful one that must make us stronger.
10:41Because it forces us to ask the right question at the right time.
10:46Not only what can AI do, but what must we continue to control?
10:51At Orange, our answer is clear.
10:54We will use AI to think better, protect better, and connect better.
11:00But we will not let it sink in our place.
11:03We will help build the infrastructure, the security, and the human framework that trustworthy AI requires.
11:10And that is why we trust the future.
