France: France 24

20260920 10:30 UTC · 00:31:00 · 447 transcript segments · GDELT Visual Explorer · plain-text transcript · Event Map

Film strip

One frame every 4 seconds, 466 in all. Click a frame to jump the transcript to that moment, or click a line of the transcript to see what was on screen while it was said. Served from apprised.news, so it loads behind proxies that block Google Cloud Storage.

00:00:00
00:00:04
00:00:08
00:00:12
00:00:16
00:00:20
00:00:24
00:00:28
00:00:32
00:00:36
00:00:40
00:00:44
00:00:48
00:00:52
00:00:56
00:01:00
00:01:04
00:01:08
00:01:12
00:01:16
00:01:20
00:01:24
00:01:28
00:01:32
00:01:36
00:01:40
00:01:44
00:01:48
00:01:52
00:01:56
00:02:00
00:02:04
00:02:08
00:02:12
00:02:16
00:02:20
00:02:24
00:02:28
00:02:32
00:02:36
00:02:40
00:02:44
00:02:48
00:02:52
00:02:56
00:03:00
00:03:04
00:03:08
00:03:12
00:03:16
00:03:20
00:03:24
00:03:28
00:03:32
00:03:36
00:03:40
00:03:44
00:03:48
00:03:52
00:03:56
00:04:00
00:04:04
00:04:08
00:04:12
00:04:16
00:04:20
00:04:24
00:04:28
00:04:32
00:04:36
00:04:40
00:04:44
00:04:48
00:04:52
00:04:56
00:05:00
00:05:04
00:05:08
00:05:12
00:05:16
00:05:20
00:05:24
00:05:28
00:05:32
00:05:36
00:05:40
00:05:44
00:05:48
00:05:52
00:05:56
00:06:00
00:06:04
00:06:08
00:06:12
00:06:16
00:06:20
00:06:24
00:06:28
00:06:32
00:06:36
00:06:40
00:06:44
00:06:48
00:06:52
00:06:56
00:07:00
00:07:04
00:07:08
00:07:12
00:07:16
00:07:20
00:07:24
00:07:28
00:07:32
00:07:36
00:07:40
00:07:44
00:07:48
00:07:52
00:07:56
00:08:00
00:08:04
00:08:08
00:08:12
00:08:16
00:08:20
00:08:24
00:08:28
00:08:32
00:08:36
00:08:40
00:08:44
00:08:48
00:08:52
00:08:56
00:09:00
00:09:04
00:09:08
00:09:12
00:09:16
00:09:20
00:09:24
00:09:28
00:09:32
00:09:36
00:09:40
00:09:44
00:09:48
00:09:52
00:09:56
00:10:00
00:10:04
00:10:08
00:10:12
00:10:16
00:10:20
00:10:24
00:10:28
00:10:32
00:10:36
00:10:40
00:10:44
00:10:48
00:10:52
00:10:56
00:11:00
00:11:04
00:11:08
00:11:12
00:11:16
00:11:20
00:11:24
00:11:28
00:11:32
00:11:36
00:11:40
00:11:44
00:11:48
00:11:52
00:11:56
00:12:00
00:12:04
00:12:08
00:12:12
00:12:16
00:12:20
00:12:24
00:12:28
00:12:32
00:12:36
00:12:40
00:12:44
00:12:48
00:12:52
00:12:56
00:13:00
00:13:04
00:13:08
00:13:12
00:13:16
00:13:20
00:13:24
00:13:28
00:13:32
00:13:36
00:13:40
00:13:44
00:13:48
00:13:52
00:13:56
00:14:00
00:14:04
00:14:08
00:14:12
00:14:16
00:14:20
00:14:24
00:14:28
00:14:32
00:14:36
00:14:40
00:14:44
00:14:48
00:14:52
00:14:56
00:15:00
00:15:04
00:15:08
00:15:12
00:15:16
00:15:20
00:15:24
00:15:28
00:15:32
00:15:36
00:15:40
00:15:44
00:15:48
00:15:52
00:15:56
00:16:00
00:16:04
00:16:08
00:16:12
00:16:16
00:16:20
00:16:24
00:16:28
00:16:32
00:16:36
00:16:40
00:16:44
00:16:48
00:16:52
00:16:56
00:17:00
00:17:04
00:17:08
00:17:12
00:17:16
00:17:20
00:17:24
00:17:28
00:17:32
00:17:36
00:17:40
00:17:44
00:17:48
00:17:52
00:17:56
00:18:00
00:18:04
00:18:08
00:18:12
00:18:16
00:18:20
00:18:24
00:18:28
00:18:32
00:18:36
00:18:40
00:18:44
00:18:48
00:18:52
00:18:56
00:19:00
00:19:04
00:19:08
00:19:12
00:19:16
00:19:20
00:19:24
00:19:28
00:19:32
00:19:36
00:19:40
00:19:44
00:19:48
00:19:52
00:19:56
00:20:00
00:20:04
00:20:08
00:20:12
00:20:16
00:20:20
00:20:24
00:20:28
00:20:32
00:20:36
00:20:40
00:20:44
00:20:48
00:20:52
00:20:56
00:21:00
00:21:04
00:21:08
00:21:12
00:21:16
00:21:20
00:21:24
00:21:28
00:21:32
00:21:36
00:21:40
00:21:44
00:21:48
00:21:52
00:21:56
00:22:00
00:22:04
00:22:08
00:22:12
00:22:16
00:22:20
00:22:24
00:22:28
00:22:32
00:22:36
00:22:40
00:22:44
00:22:48
00:22:52
00:22:56
00:23:00
00:23:04
00:23:08
00:23:12
00:23:16
00:23:20
00:23:24
00:23:28
00:23:32
00:23:36
00:23:40
00:23:44
00:23:48
00:23:52
00:23:56
00:24:00
00:24:04
00:24:08
00:24:12
00:24:16
00:24:20
00:24:24
00:24:28
00:24:32
00:24:36
00:24:40
00:24:44
00:24:48
00:24:52
00:24:56
00:25:00
00:25:04
00:25:08
00:25:12
00:25:16
00:25:20
00:25:24
00:25:28
00:25:32
00:25:36
00:25:40
00:25:44
00:25:48
00:25:52
00:25:56
00:26:00
00:26:04
00:26:08
00:26:12
00:26:16
00:26:20
00:26:24
00:26:28
00:26:32
00:26:36
00:26:40
00:26:44
00:26:48
00:26:52
00:26:56
00:27:00
00:27:04
00:27:08
00:27:12
00:27:16
00:27:20
00:27:24
00:27:28
00:27:32
00:27:36
00:27:40
00:27:44
00:27:48
00:27:52
00:27:56
00:28:00
00:28:04
00:28:08
00:28:12
00:28:16
00:28:20
00:28:24
00:28:28
00:28:32
00:28:36
00:28:40
00:28:44
00:28:48
00:28:52
00:28:56
00:29:00
00:29:04
00:29:08
00:29:12
00:29:16
00:29:20
00:29:24
00:29:28
00:29:32
00:29:36
00:29:40
00:29:44
00:29:48
00:29:52
00:29:56
00:30:00
00:30:04
00:30:08
00:30:12
00:30:16
00:30:20
00:30:24
00:30:28
00:30:32
00:30:36
00:30:40
00:30:44
00:30:48
00:30:52
00:30:56
00:31:00

Transcript

Google Cloud Speech-to-Text API (Chirp) + Gemini 2.5 Flash Non-Thinking. Treat it as a searchable index of what was broadcast, not a quotation record.

00:00:00that they are possibly acting in good faith
00:00:01the main leading laboratories to or are they fueling a kind of fear marketing?
00:00:05And all of this, well, it's up to us to be vigilant somewhere. This
00:00:08fear marketing, as has been said previously, is something that
00:00:12can be favorable at a given time, which is
00:00:15precisely the one we are currently experiencing in the generative artificial intelligence markets.
00:00:19That is to say, when the actors move from what
00:00:23we call an emerging market phase to enter a growth phase, and where it is a matter
00:00:27of securing the investments that have been made in this
00:00:30sector in order to be able to appear sustainably as
00:00:34actors who are capable of dominating the competition and eliminating the
00:00:37others. Then very often we see actors
00:00:42different logic of AI in these markets calling for more regulation and very often they have to resort to
00:00:44fear marketing by saying somewhere regulate me because
00:00:48if you don't regulate me we will face
00:00:51big catastrophes and obviously when they take this
00:00:54approach with the
00:00:56speech, her State of the Union speech, regulator they try to incite the regulator to put in place
00:01:00mechanisms that are favorable to the technology, to the standardization of the
00:01:03technology on which they have bet a lot and when we know that
00:01:07today indeed there is a big competition on two very different visions
00:01:11both technologically and ultimately on what can be done with
00:01:14artificial intelligence between what is the taste between two imperialist blocs, which is the American
00:01:18bloc on one side and the Chinese bloc on the other
00:01:21hand, who have made different technological choices and who consider very
00:01:25differently because at the American level we are really on
00:01:29the edge of... border where at the Chinese level we are
00:01:32more on it as a kind of common good and well if I put
00:01:36in place a regulation that is favorable to my technological
00:01:39choices, not only do I protect myself but I also
00:01:44the phenomenal amount of exclude de facto actors who have made other technological choices and notably the Chinese
00:01:47IPO at this point in actors, so indeed this fear marketing can be
00:01:50savagely used for this purpose. These are therefore giants who pretend to be afraid
00:01:54because they know how to pilot drones with to call for tailor-made regulation
00:01:57to block any form of competition. Well,
00:02:00personally, I don't agree with this interpretation. Both
00:02:04Dario Amodey, the CEO of Anthropic
00:02:07and Sam Altman, the one from Open, had
00:02:11mentioned years before founding their current company that the
00:02:15greatest danger humanity faced was the development of AI and
00:02:18uh when we talk about
00:02:23that artificial intelligence when we talk about these developments uh I don't think it really
00:02:26makes sense for them to slow down. Anthropic is the company that...
00:02:30in the history of capitalism has grown the fastest, they have grown from
00:02:33a revenue of 10 billion at the beginning of the year to 60 billion, it's unprecedented, but it took
00:02:37years and years to do the equivalent, so I don't think they
00:02:41have an interest in slowing down their
00:02:43growth. He says
00:02:46precisely
00:02:48they pretend to want to stop and ask for tailor-made regulation
00:02:51for but if it slows them down it slows down growth
00:02:57no, I'm sorry, excuse me, but I wanted to tell you anyway that I'm afraid. But I'm not afraid of
00:03:00AI, I'm afraid of what Mrs. von der Leyen declared, we'll come to Mrs.
00:03:04von der, that's in the 3rd part, I'm afraid that we don't have AI in Europe and I think
00:03:08that's what they are doing today, keeping our heads underwater by installing a
00:03:12governance piloted by the United States which will hinder the
00:03:15development of AI in other countries of the world and by the way, it's we can draw a parallel with
00:03:18sustainable development, for a long time the developed countries kept the heads of
00:03:22underdeveloped countries underwater with environmental
00:03:25regulations and the countries of the south said but let us grow, it's a
00:03:29debate we've seen... for almost 30 years and well we're going to witness the same debate today, except that today the
00:03:33underdeveloped countries in terms of AI are us and so I think the real subject
00:03:36of this debate is not American because we are talking about an
00:03:40economic and political debate in the United States, it is a global
00:03:43debate on global governance, perhaps tomorrow we need to
00:03:46imagine for AI and certainly not American. No,
00:03:50it's good for you to react, yes, but in the end,
00:03:53uh, you never talk about the people, about the citizens who watch this. I started
00:03:57with that, I talked about techno-shock, I talked about civilization.
00:04:00Excuse me, I think the real subject, however, is
00:04:03how we are going to protect people and make them more intelligent in the face of these machines, we must
00:04:07educate, who talked about education there, no one, we talked about regulation.
00:04:11I think we have a card to play anyway, because the Chinese and the Americans
00:04:14slow down and somehow don't agree on what needs to be done, it's a cold
00:04:18war and we will need counter-powers, we will need people capable
00:04:21of evaluating and having AI tools perhaps for
00:04:24security, before perhaps talking about this clash between China and the United States and
00:04:29behind, most of the then about Europe. Sam Altman proposes a system for reporting incidents inspired by
00:04:32aviation with
00:04:34Wednesday in her State of the Union mechanisms comparable to those of
00:04:36voilà, agents who are supposed to verify this in the
00:04:40United States, is it the right way? It's a good way, but it must
00:04:43be external people because they themselves said at the same time
00:04:47West, China has about yes that he proposed their own
00:04:50tool. From the beginning I've been ethical, excuse me, you
00:04:54are completely naive, madam, the public affairs and communications departments
00:04:58of these companies employ thousands
00:05:00of people, among the best lawyers in the world, the best
00:05:02paid.
00:05:06I think that this regulation that is proposed
00:05:09warning, AI has an enormous part of factualism in this approach. I think this regulation is
00:05:13useful, that's what perhaps differentiates us, not if it comes from them, but
00:05:17that's what I'm saying too, we must not let, however,
00:05:20we must not let the Americans propose their own
00:05:23regulation, otherwise what do they do, the tools, the regulation and the insurance,
00:05:27they will manage everything, so that's certainly not what should be done, they have China opposite them who disagrees
00:05:31who will fight against them, that's where we perhaps have a role to play, saying attention, we need
00:05:35an outsider, they asked for outsiders,
00:05:38the Americans proposed, Altman and Dario asked
00:05:44exactly and the problem they have is that currently the administration is completely
00:05:47anti-regulation and I will propose a summary point between my two
00:05:51colleagues which is that I completely agree that the subject of Europe is
00:05:55extremely important, Europe must accelerate and the reason why Europe does not appear
00:05:58in the speeches of Dar mode of Samman is that Europe currently
00:06:02is years behind, we will
00:06:05talk about it, it's simply not finished, but Europe has a regulation, absolutely, but don't
00:06:08worry, we will talk about it, we have time to talk about it, I just want to tell you that
00:06:12artificial intelligence also feeds an immense
00:06:16financial speculation, valuations are soaring, investments are
00:06:20multiplying, announcements are coming one after another, OpenAI
00:06:23also announced last weekend that it had postponed its
00:06:26IPO, Julien Pillaot, was the
00:06:30moment
00:06:31ill-chosen?
00:06:35Is there a good time to launch an IPO? If
00:06:39we knew, if we could demonstrate it scientifically.
00:06:43Uh, the fact remains that indeed at OpenAI today there is still a problem
00:06:46which is that they cannot demonstrate the profitability of their
00:06:50models, while Anthropic has managed to reach this profitability
00:06:53threshold in the last two quarters, so
00:06:57launching an
00:07:00history when you have your main competitor
00:07:03today who is preparing his IPO
00:07:06at an interval and who, he exact, can have an
00:07:09IPO, thank you Karina for correcting me. Uh,
00:07:13but he, well, has fundamentals that seem, from the investors'
00:07:16point of view, perhaps a little more solid, this would be a little bit subsidiary and
00:07:20so OpenAI made the decision, it seems to me, quite
00:07:24reasonable to postpone this IPO to a little later
00:07:27in order to perhaps make up for its delay on the path to profitability. Nevertheless,
00:07:31the investments
00:07:35companies, and especially what we call the Magnificent Seven,
00:07:39these are the seven biggest American artificial intelligence companies, continue
00:07:43to increase year after year to reach dizzying amounts. We will
00:07:46probably be at the end of 2026 at more than
00:07:49750 billion of cumulative investments in artificial intelligence in
00:07:53infrastructure and model training, it's
00:07:56colossal and it can outline the contours of a bubble about
00:08:00which we can ask ourselves whether it can burst or whether it will perhaps deflate in the coming months
00:08:04or years and what is your diagnosis on that? Will this bubble burst
00:08:08Julien Piot?
00:08:10There, similarly, if I were tomorrow that would be and
00:08:14probably I would take stock market orders that would allow me to bet
00:08:18on the future that I do not know, but no
00:08:21no, joking aside, obviously, uh, my
00:08:24current diagnosis is that there are obviously
00:08:28risk factors for the bursting of the bubble and notably the lack of profitability of many
00:08:32models among actors who are specialized in generative artificial intelligence who have
00:08:35difficulty making their models profitable, but there are also factors
00:08:38that make this bubble quite robust
00:08:42in fact to be able to somehow resist for a certain time and
00:08:46among these factors there is one that I particularly follow which
00:08:49is the circularity of investments in artificial
00:08:53intelligence. Behind this somewhat technical term, what is there? There is the idea that
00:08:56all these major artificial intelligence actors are all
00:09:00interdependent because they all have either stakes in the
00:09:03companies or cross-investments with the
00:09:06companies or supplier-customer partnerships with the
00:09:10companies and so there we could see actors like
00:09:14Google, like Microsoft, like OpenAI, like Nvidia, as being
00:09:17competitors, in truth, they are rather competitors, these
00:09:21are companies that compete on certain market segments, but in
00:09:25truth they are all jointly engaged in a kind of community of
00:09:28destiny, because they have cross-investments, and these
00:09:32companies, which are the ones that have the most interest in artificial
00:09:35intelligence becoming profitable and sustainably profitable, are also
00:09:39the ones that are most interested today in the valuations of the companies they
00:09:43own also increasing, so for the bubble to deflate, these actors
00:09:47would ultimately, somehow, no longer believe in it themselves, and there is
00:09:50little chance of that. This
00:09:53interdependence is very improbable, do you agree? It prevents
00:09:56the bursting of this bubble. It sells these chips to
00:10:00people who make models, so they have to make models to buy the chips, basically it's a loop like
00:10:03that. On the other hand, for the gold rush, it was those who made the picks and shovels who won,
00:10:07so Nvidia will always win, the others will pass and now
00:10:11they have reached a size that... too big to fail, as they say, what we said
00:10:14at the time of the subprime, of the real estate loans, the bubble had
00:10:18burst at the time, if they burst there, it's the whole American economy
00:10:21that collapses, the 2 % of American growth. Two of the
00:10:25percent of American growth are the data centers invested in the
00:10:29United States, there are still two points on which I wanted to draw your attention, I agree with what has just been
00:10:33said, the first is the extraordinary hubris of these bosses, they have
00:10:36created brilliant companies in a very short time from nothing, that incites a form of
00:10:40madness I think, in any case a surpassing of oneself and it's quite human, I can
00:10:44understand it and it has been seen in the history of economics a thousand times from Rockefeller
00:10:47to uh Steve Jobs, well there you go. And the
00:10:51second point is that the two phenomena that were predicted to us in 2022,
00:10:55massive job destruction and productivity gains that were supposed to
00:10:58happen very early for jobs and a little more recently for
00:11:02productivity, did not happen. So what major groups observe, we can
00:11:06think so, but for now no, what major groups who are clients of these AI solutions
00:11:10and notably of those from Anthropic observe is that they haven't, they have pilot projects, they have tested
00:11:14deployments, they have created new products but they have not had the expected
00:11:17productivity gains. This I think is a major parameter, this
00:11:21might change with robotics that is advancing full throttle. I have done a lot of
00:11:24robotics. Affective, social, etc., the Chinese show great
00:11:28strength, you see people on TikTok folding laundry or doing
00:11:31sewing machine robots, there are robots that will replace workers, that's where
00:11:35the blue-collar school will build them for textiles for example and it's in India and
00:11:39China, they will have a flood, we are talking about 70% at the world forum,
00:11:43but for how long, how long will it take to have a few
00:11:46years, by then the bubble, by
00:11:49then you no longer have, you have a bubble necessarily in the middle because these investments. except
00:11:53that there is a moment of market clarification before the arrival of robotics, that's
00:11:57it, that's my feeling. but you want
00:11:59to react?
00:12:01No, indeed, just to add a small
00:12:04nuance to what has just been said, do we realize
00:12:09capital needed, of materials needed and energy needed to
00:12:12replace all human labor, blue-collar or white-collar, with
00:12:16machines? So indeed, at some point, there will still be conflicts
00:12:19of use that will appear and will slow down
00:12:22the progression of this phenomenon that
00:12:25has been described, that's it. To what extent will it
00:12:28slow down? The story remains to be written, but it's true that there are always forces that pull
00:12:31back, and those who predicted the
00:12:34replacement of one in two workers just
00:12:383 years ago when generative intelligence appeared in
00:12:42certain sectors today can look back at their words and say, "Well,
00:12:45in truth, it didn't work as expected," simply because
00:12:49the productivity gains weren't there, because there are forces pulling back in the market that make it
00:12:53not necessarily possible, and because it turned out to be much more expensive than we
00:12:57thought in use, or there's more to it than that, because we're not going to replace tasks,
00:13:01so we saw very well, for example, with Geoffrey Hinton, a Turing
00:13:04Award winner, who said 6 years ago, a mathematics prize in
00:13:08physics, in physics perhaps, to clarify, because the Turing Award
00:13:12is a computer science award, but there is also the Nobel Prize in
00:13:14physics, and he had predicted 10 years ago, in
00:13:182016, that radiologists would disappear from the planet because
00:13:21we knew how to detect early cancers on X-rays.
00:13:25However, the opposite happened, not at all, there is even a shortage of
00:13:29radiologists in the United States, it is a profession where they have been able to use these
00:13:33tools in their expertise, because it was just a task,
00:13:36it was reading X-rays, but when you are a radiologist, you welcome
00:13:39patients, evaluate what the machine says, whether it is correct or not, re-examine
00:13:43other images and so it is much more and it is perhaps 1/5th
00:13:47of their profession and so in the profession it will be necessary to understand for everything that
00:13:51concerns everything that is about white-collar and gray-collar.
00:13:55It will be necessary to understand the limit well, that is, not to replace
00:13:58humans because we need humans in the loop and not just for security
00:14:01problems, and that is the human-machine relationship
00:14:05and I think we don't work enough on that, in any case, it's said and thank you for this clarification,
00:14:09meanwhile, Donald Trump, he wants to continue, he wants the
00:14:13United States, as you said earlier, to maintain its lead.
00:14:17Uh, can American AI giants really slow
00:14:20down in this competition, notably
00:14:23uh with China?
00:14:26One, there are two problems from the point of view of companies, there is a legal
00:14:30problem that companies in the United States
00:14:33are subject to antitrust laws which are
00:14:36anti-competition laws that mean they cannot
00:14:40voluntarily and legally decide to coordinate to
00:14:44slow down, that's why they asked for American intervention, and then there is
00:14:47obviously the subject of China. The subject of
00:14:50China is delicate because China has extremely high engineering
00:14:53skills and very great manufacturing
00:14:57skills, but China still has two very
00:15:00big disadvantages, the first is that of export
00:15:04controls, namely that the computing units called
00:15:07GPUs that are mostly developed in the
00:15:1220 times less if we count in performance than the
00:15:15United States, and so they are still playing with a very very big disadvantage and
00:15:19the second is that China currently is a little bit
00:15:22uh the image that... I like to keep in mind is
00:15:26the United States is the tow truck and China is the car
00:15:29behind, in the sense that it is much easier
00:15:32to develop models by copying a little
00:15:35bit, by challenging. Not being a leader exactly, using existing
00:15:39models to develop a model that is
00:15:42similar, it comes from a mechanism called distillation, which is the
00:15:45training of a model on the results
00:15:49of another model, so you use the results of ChatGPT to train a model
00:15:53that approaches this competence, I think what
00:15:57you say is very true, but we need to put it simply, uh, it's a
00:16:00race for AI between China and the
00:16:03United States, Europe does not exist. Mensch also says not
00:16:06to. Here, we are talking about
00:16:08China. China absolutely wants to win this
00:16:12battle, it is determined to do it with open source models
00:16:16that paralyze the Americans because it is a
00:16:19distribution, that's what it is, which is very
00:16:23aggressive, uh, they have a different way of functioning, they
00:16:26do not have this scrutiny from the media and markets
00:16:29on their developments, they can offer AI free of charge to allied
00:16:33countries, thus forming coalitions, uh, so I think it's a battle, uh,
00:16:37uh, I want to say definitive and ultimate, because if there's one
00:16:41use of AI that doesn't suffer from dispute, it's war. Today,
00:16:44Thales has just announced that it is releasing a competing system to Palantir's
00:16:48for massive targeting of targets, so to go from 100 targets to 100
00:16:52targeted targets in a day, that's what Palantir allows the
00:16:56American army today. NATO has just become a client of Palantir. The question tomorrow is
00:17:00who will have the best weapon systems to pilot their drones, their planes,
00:17:04etc. So it's a battle to the death, I want to say, between the two blocs. It's a
00:17:07cold war, you have a very small problem too anyway, which is that open source allows for
00:17:11enormous criminality. So not only will it
00:17:14help with war and they will beat everyone and China, that
00:17:18gives the information because you can do anything with that. There are even
00:17:22now warriors, whether they are in Russia or Ukraine, who sell
00:17:26themselves for a lot of money in Mexico to go help drug cartels
00:17:31AI. We will find that everywhere in the hot spots of the planet
00:17:35with mercenaries capable of piloting that. So all the hackers,
00:17:39really, by leaving these objects open, I think we are creating
00:17:42a serious problem and China will have this problem.
00:17:46If Donald Trump and Xi Jinping are to meet at the end of the
00:17:49month, we imagine that artificial intelligence should
00:17:53be on the agenda of their discussion. What can we expect from this face-to-face?
00:17:56What is delicate is that on Trump's side he announced very
00:18:00clearly that he really didn't want
00:18:03to slow down AI so
00:18:06he will probably... try to do as
00:18:09little as possible. Uh on the Chinese side too, they
00:18:12think for them it's an, as my
00:18:16colleague said, it's existential and if
00:18:18Anthropic or OpenAI are first, they see it as an existential
00:18:22threat to their national sovereignty and therefore
00:18:25uh my hope is that there will be
00:18:28discussions on these issues of potentially
00:18:31slightly slowing down the development but I don't expect as much from
00:18:35it quickly. There are signs that show that the
00:18:39Chinese president is adopting the verbiage of Anthropic and so on, saying
00:18:43that he would like to have external
00:18:46committees to verify what AI is doing, which is still
00:18:49a sign of listening. Julien Pillot, I hear you
00:18:53react very briefly, please. Briefly, no, but today there is little
00:18:56reason to think that
00:19:00investments and the AI race can slow down when you have two imperialist
00:19:03blocs clashing like the United States and China.
00:19:07Simply because it outlines the game of a prisoner's dilemma where
00:19:11neither has an interest in slowing down if they are not certain that
00:19:14the other will also slow down, at the risk of shooting themselves in the
00:19:18foot, and where there is a second phenomenon that is added on top, called the tragedy
00:19:22of horizons, that is to say that currently, and it has been
00:19:25said earlier, investments in AI support
00:19:28American growth and to a lesser extent Chinese growth, so these are
00:19:32important and short-term gains that are certain and the risks that
00:19:36exist are more distant, vague,
00:19:39uncertain, so when you combine these two reasons,
00:19:43there is little hope that the states, whether they are
00:19:46the United States or China, will ask their AI champions to
00:19:52stop the fun, but it's very important that they talk about it and we start a debate
00:19:56on it, because Europe has something to play in this
00:19:59diplomacy between these two great states that do not want to
00:20:02regulate. So, precisely, Europe, as we said, is
00:20:06AI models are... designed by the United States and China,
00:20:11address, the President of the European Commission once again insisted on
00:20:15the regulation of artificial intelligence,
00:20:17listen:
00:20:21the AI Act is of course a crucial element for putting in place
00:20:25safeguards, and it puts Europe in a position to shape global
00:20:28efforts to address this challenge. We will therefore
00:20:32intensify our efforts alongside our like-minded partners such
00:20:35as Canada, the United Kingdom,
00:20:38and others. We want to join forces on
00:20:42model evaluation, verification, early
00:20:47security, and much more, and I would invite
00:20:51discuss how we can support current industry efforts
00:20:55to moderate the race for
00:20:57innovation,
00:21:02so everyone wants to react very briefly, no I find these comments
00:21:04inadmissible,
00:21:07she represents us, she is Europe, she should say exactly the opposite, she should
00:21:11say our regulations have not worked, the AI Act has been half
00:21:14applied, there is a consensus to say that today it is no longer relevant, it has
00:21:1811 directives and 50 rules. the European Union on AI and data protection,
00:21:21of which a good half are obsolete and the other half are blocking our businesses
00:21:25on the ground, she should say: my God, let's catch up, and then she tells us
00:21:29we need to slow down because maybe it presents a risk,
00:21:32honestly, I find that the call for regulation and
00:21:36caution and deceleration and countering with the
00:21:39countries that shared our values, she didn't say we were decelerating I think, she said we need to
00:21:43curb the development of AI, it's in her
00:21:46but she has no power over what others decide to do. We
00:21:50can distinguish two elements, which are the frontier and Europe, Europe is so far behind the
00:21:53frontier that indeed we need to accelerate the 600 billion in capital, I don't know
00:21:57how much it is but it's 10 billion maybe, and so Europe clearly needs to
00:22:00accelerate but also I think that we need a
00:22:04discourse and at the same time where
00:22:07we both move forward, we accelerate in Europe, we make
00:22:10better laws on what allows us to
00:22:13build data centers, so for example
00:22:16permitting, building permits, etc.
00:22:20we allow financial engineering similar to what has been done in the
00:22:24United States and at the same time we need to have an appropriate discourse on
00:22:28the risks on the frontier and on regulation. Laurence de Viller, is Europe
00:22:32in a position to shape global regulatory efforts, can
00:22:36we regulate a technology that we don't produce for now, no we don't know
00:22:40how to do it, we need research and research is the poor relative, so the economy wants to play
00:22:43all by itself, the day when AI will be even smarter and no one will be able to
00:22:47know what is really happening, we will be well advanced. We need measures for
00:22:51evaluation, for example I have PhD students who work on benchmarks,
00:22:54these are test sets and on which
00:22:58we test AI and often it has already ingested it,
00:23:01so when it answers it's not by reasoning but it's by copy-pasting in its memory somewhere,
00:23:04so we need measures everywhere to effectively
00:23:08find where the real pitfalls are and manage to
00:23:12evaluate correctly, we need audits, we need insurers and it shouldn't
00:23:15only be the United States doing this and that we are also in the loop. Julien
00:23:19Pilot, does Europe impose its standards on the rest of the world as it
00:23:22did for example with data
00:23:24protection?
00:23:28In the case of AI, it will be complicated because indeed, to be able to impose what
00:23:32is called a "Brussels effect", that is to say to impose normative power on the rest
00:23:35of the world, two things are needed: a market that is absolutely essential
00:23:39for us to use the depth of this
00:23:42market to pressure foreign actors to adopt the standards
00:23:46and comply with the notes, we are still a little bit and then also we need to...
00:23:50see beyond beautiful laws, beautiful rules, the ability to enforce them by
00:23:53engaging in a power struggle with the most powerful states and/or companies
00:23:57of the states in question, and these last few years have rather shown some signs
00:24:00of weakness at this level because in fact we are extremely dependent on what
00:24:04the states or companies we are talking about can provide us elsewhere, if only
00:24:08to think of Starlink today which has become absolutely essential on the Ukrainian
00:24:12front line, well that somewhat slows down the desires we
00:24:16might have regarding Elon Musk on certain matters and
00:24:19certain. compliance issues that certain matters could have, especially regarding
00:24:22the DSA for X, so here we are, today we are a bit in a weak situation
00:24:26because our dependency schemes are such technologically that we are afraid
00:24:30in truth to follow what is called a kill switch, that is, that
00:24:33foreign actors deprive us of access to
00:24:36certain services that have become absolutely essential for us, yes in our
00:24:40daily lives, listen, that will be the last word, thank you very much to all four of you,
00:24:44Laurence Devillaire, professor in artificial intelligence and ethics at the Sorbonne, Julien
00:24:48Pilot, you were remotely. researcher at linesc
00:24:52Edouard Filali, co-founder and CEO of Jean, communication agency and
00:24:56digital strategy specialist, Simeon Campos,
00:24:59you are an entrepreneur in AI, thank you for following us,
00:25:03the news continues on France 24.
00:29:33They are Kurdish people who live between fear and uncertainty,
00:29:36waiting for
00:29:38the unknown.
00:29:41has been organized to give them clothes.
00:30:07It's 1 p.m. in Paris, welcome, we are together until 2 p.m., you
00:30:11are watching Paris Direct, I had a doubt, we will talk about Germany in
00:30:14this edition, two regions are voting today, the stakes are such that the Chancellor
00:30:18canceled his departure for New York, he called an emergency
00:30:21meeting, the AFD could make new gains after its
00:30:24victory in Saxony, even if the far left is very mobilized in
00:30:28Berlin. Facing the latest developments in Yemen, Donald Trump
00:30:31cut short his weekend. The Houthis are increasing pressure on Saudi Arabia and
00:30:35are advancing eastward, how far do they want to go, what is the kingdom's
00:30:38ability to stem their advance, we will discuss this with David
00:30:42Rigolet Rose who will be with us in a
00:30:44moment.
00:30:46It's effervescence in Paris and not only, 5 days before the Pope's
00:30:50arrival, the enthusiasm is strong, it was necessary to extend the security perimeter for
00:30:53the open-air mass that will be celebrated on the Place de la Concorde,
00:30:5760,000 people are expected.
Data courtesy of The GDELT Project (gdeltproject.org), from the Internet Archive TV News Archive. Film strip and transcript are GDELT's, rehosted here under their terms of use, which permit it with this citation.