The Paris AI summit is over and it is time to try to draw the bottom line and even sketch out possible scenarios for the future. Indeed, we must not delude ourselves: whatever the intentions of the summit’s instigators, AI was not the subject but a means. A means of establishing technological leadership from which domination and economic subjugation would flow.
But was there any point in expecting anything else? During a soccer World Cup, people are interested in the teams, not the ball (except when it goes out of bounds).
Massive investment in AI in Europe
There has been talk of a world summit, yet most of the investment announcements have been for Europe. Surprising? Not at all. In a race, it is the stragglers who have to pick up the pace massively, not those who are already ahead and have opened up a gap.
200 billion euros of investment (public and private) have therefore been announced by the European Commission for the European Union as a whole and 109 billion euros of private investment from France and abroad in France for the coming years ([FR]At the AI Summit, announcements of massive investment… and a step backwards).
This covers a wide range of areas, including support for the ecosystem, the funding of research projects, the creation of infrastructures and data centers, the development of sovereign open source AI, the training of talent, the retraining of workers affected by AI with the identification of priority sectors such as health (AI for medical diagnosis), energy transition (optimization of electrical networks), cybersecurity and autonomous transportation.
It is not known whether this will enable Europe to catch up, but France is clearly positioning itself to be at the head of the chasing pack, hoping to be the driving force that will enable it to close the gap with the leaders. And, obviously, its partners, both European and non-European, subscribe to and endorse this idea.
International investment but different motivations
One may wonder how a financially drained Europe can afford such announcements, even if, in terms of expenditure, investment and operating expenses should never be lumped together.
The answer is simple: a very large part of the investment will be international, which may seem contradictory to the idea of sovereignty underlying the “AI war”.
Thus, among the investors are the United Arab Emirates (50 billion euros) for a data center and a campus in France, the Canadian fund Brookfield (20 billion euros) for data centers and infrastructure, Amazon (6 billion euros) for cloud infrastructure, the American fund Appolo (5 billion euros) for energy connectivity and, last but not least, Iliad (€3 billion), Mistral AI (“several billion”) and Bpifrance (€10 billion).
If everyone is investing in the same place, it is not, in my opinion, for the same reasons.
The first is certainly shared and economic in nature: it is easier to get on a train at the platform than on a movingtrain, and the Americans and Chinese are not waiting for anyone, especially since the time is rather for protectionism.
The second is geopolitical. After the Second World War, there were two blocs, East and West, and the non-aligned. Here the blocs are no longer regional but national (USA and China) and Europe, if it would one day like to form the third bloc, is more, with France in the lead, the figurehead of the non-aligned. And since there is expertise and potentially fertile ground there, those with money see it as the right place to develop the technologies that will enable them to avoid dependence on the two giants. In this category I include of course the Emirates but also Canada, which has understood that its imposing neighbor will not always be an ally.
The third is, again, economic in nature, and here I am thinking of Appolo, Amazon and even Brookfield. It is a question of diversifying or even derisking their investments. If the European project were to work, it would be a shame not to be at the table and reap the rewards, and for them this is not incompatible with being very interested in the success of the American vision. After all, on a misunderstanding they can win on both sides, and investing “only” a few billion for the first two is a trivial stake compared to the potential gain.
Ethics: between a bone of contention and backtracking
While, on the European side at least, the announcements about investment could only be welcomed as a clear positive signal, the real issue was that of ethics.
In fact, the financial aspect was of interest only to those Europeans who have a delay to make up and finance, and the real issue, the adults’ issue, was to know whether global governance of AI would contribute to leveling the differences or, on the contrary, whether we would enter into a logic of irreconcilable blocs.
It was easier to agree on a declaration than on agreements following the recent COPs on the climate. It was signed by 61 countries, including France, India and China, but not the United States and the United Kingdom.
One may well wonder what such a charter is worth without the USA (sorry for the United Kingdom, whose weight in this story is quite anecdotal) and one may well wonder what China and India are doing in this and whether there is not a risk of internal dissension appearing more quickly than one might think.
The main objective of this declaration is to promote an AI that is “open”, “inclusive” and “ethical”, while strengthening global coordination for its governance. It also aims to avoid market concentration in order to make AI more accessible and sustainable for people and the planet. Its guiding principles emphasize the accessibility of AI to reduce the digital divide, its transparency, security and reliability, as well as respect for international frameworks. It also wants to encourage innovation while avoiding market concentration, and aims to have a positive impact on the future of work and sustainable growth.
Everything the United States doesn’t want… but not only them.
A few days before the summit, 70 European companies had published a report on their ambitions in the field of AI and what they expect from the European institutions to help them (Is the ambition of European businesses in AI realistic?). A report strangely backed by an American fund and in which the word ethics appeared only once in 56 pages, as if it were being paid lip service at the end of the introduction in order to dismiss the subject.
I was about to write that these businesses had not been heard when a news item broke, but I think it is more the remarkable performance of US Vice-President JD Vance than the business lobby that is the cause of this, which also says a lot about the philosophy of EU leaders.
We are talking here about the European Commission’s decision to withdraw the directive on liability for AI from the 2025 agenda ([FR]After criticism from JD Vance, the Commission withdraws the directive on liability for AI). In the event of damage due to AI, it did not require the injured party to prove the liability of the AI supplier, as is the case under common law, but rather the supplier to prove that it was not liable.
This can be seen as a U-turn intended to reassure AI stakeholders and investors and counter the image of Europe as over-regulating, but it would be going a bit too far.
Indeed, the European AI Act (The European AI Act for Dummies) made it less urgent and the real question now is whether, for the sake of consistency, its already gradual entry into force will be delayed or even partially stripped of its substance in the near future.
But while on the one hand the declaration is in line with the wishes of society and users/consumers, this move by the Commission is openly pro business and pro tech.
Investments in Europe must not be a cover-up
If, on the European side (because let’s face it, more than a world summit it was more like a match between Europe and the rest of the world, or almost the rest of the world), we can boast in the aftermath of the summit of having obtained massive investments and made almost everyone adopt an ethical and sustainable vision of AI, we should not believe that everything has been resolved and that tomorrow we will be able to look the Americans and the Chinese in the eye.
Beyond the announcements, questions are already being asked as to who will do what, within what time frame, whether the international nature of the investments is not a threat in terms of sovereignty, but that is not all.
I find the report I mentioned above quite lucid with regard to Europe’s strengths in AI but also to the institutional and legal obstacles that curb its competitiveness, and apart from the end of underinvestment and backpedaling on the liability directive, I see nothing to tell me that businesses have been heard.
The IA Act is still seen as a problem and, in general, over-regulation can be a real obstacle to the development of local players. Investing hundreds of billions is good, but if it is to immediately handicape those who benefit from it, it will not get us very far.
This was not a European summit, so initially we did not expect anything specific on the subject, but the EU will not be able to avoid a spring clean to accompany its efforts in AI.
Is Europe at risk of becoming the world’s factory?
In a very simplistic way, AI is three things: software foundations, infrastructure and products designed on top of all that.
I have the impression, since the DeepSeek episode, that there is a shift in the importance given so far to foundations and infrastructure towards the product, since it appears (with all due caution) that we can ultimately do well in this area with more frugality.
Even if the announcements have yet to be clarified, it seems to me that everything is mainly geared towards infrastructure and I don’t see much being done to create or develop national champions, which are the prerequisite for a possible Airbus of AI.
So of course we will have the infrastructure, but to run the products created by others and become the factory of those who make the products and capture the value of the market?
Note that when we see that in France we injected 200 billion into tech to ultimately create 2.5 billion in wealth on the stock market, it may be in our interest to specialize in data centers ([FR]Should we flee France to make a successful startup?).
Many questions remain unanswered
It is clear that investment in Europe and a joint declaration on ethics were at the heart of this summit, and in this sense it can be said that it has achieved its objectives. At least for Europeans and a few others.
But AI raises many other major societal issues that could be a time bomb if we don’t take immediate action (The challenges posed by AI are not technological, but must be met today.). In this sense, this summit was not the expected “AI COP” because many topics were raised but quickly relegated to the background, even though it will only accelerate their scale and speed of appearance (Like the climate, does AI deserve its “COP”?).
Impact on employment, environmental impact ([FR]“AI and ecological disruption: getting away from the techno-solutionist illusion”, copyright, geopolitical risks… move along, there’s nothing to see, when in fact the summit has created a time bomb on each of these subjects.
Bottom line
We can talk about investments, ethics, current and future societal problems, and there is an infinite number of things to say on the subject, but to focus on all the collaborations and convergences would perhaps be to miss the most fundamental element of this summit.
Just as the Berlin blockade in the aftermath of the Second World War formalized the division of the world into two ideological blocs, the AI summit formalized much the same thing, although it is not yet clear whether the United States is pitted against the rest of the world, or whether there are three blocs made up of the USA, China and the non-aligned countries, with Europe in the lead.
But it is this divide that is likely to define the world for decades to come, with AI being just one of the playing fields.
Image : AI Investment de TSViPhoto via Shutterstock.
