Gilles Babinet, 56, likes to quote his tech companions. Silicon Valley leaders to the champions of the French ecosystem, passing by the former director of the National Agency for the Security of Information Systems (Anssi), Guillaume Poupard, to a handful of nerds like him keen on foresight, to the image of the transhumanist Laurent Alexandre with whom he loves to scrap. It was again thanks to a childhood “buddy”, whose dad had bought one of the first computers in circulation, that the little seed of computing germinated in young Gilles, he says. “A real stroke of luck.”
This thick catalog of encounters and acquaintances explains the many hats that this self-taught fifty-year-old wears today. Non-exhaustive list: successful entrepreneur (Musiwave, sold for a hundred million dollars), president of the National Digital Council (CNNum), “digital champion” of France with the European Commission, teacher at Science Po Paris, or still a writer. Gilles Babinet is the author of six books, the most recent of which, How Hippies, God, and Science Invented the Internet (Odile Jacob), published in May, takes the reader on a fascinating epistemological journey. The ideal starting point to approach the future of the Web, at the time of ChatGPT and generative artificial intelligences (AI) called to upset our lives.
“We live in a reactionary internet age”
Gilles Babinet’s observation is simple: major ideological orientations shape our technologies. And not the reverse. The former, in a “military, Cold War” context, were sovereign. Some will say: secret. “We don’t really know what was going on at the Bletchley Park site in the UK, for example [NDLR : où le mathématicien Alan Turing déchiffrait les messages des nazis]”, he illustrates. This era was followed by a new desire for transparency, leading in particular to the publication of the TCP/IP protocol in the 1970s, used for transferring data on the Internet. Babinet met the American Vinton Gray Cerf, one of its inventors, who confirmed to him his fear of seeing, at the time, his work “monopolized by the military”.
This rejection of secrecy was then part of an intellectual context across the Atlantic turned towards the well-being of humanity, and fueled by “science fiction, the idea that information should be open, the challenge to the regard to the SRI, the Stanford Research Institute which had developed for the US Air Force the ancestor of the Internet, the Arpanet, or even the Hippie movement”. A narrative recovered later by startupers with a more “commercial” spirit.
What characterizes the Internet generation in which we live? A more reactionary approach, a “backlash from globalization”, according to Babinet. “We have seen the emergence of technologies developed against the States, like those financed by powerful investors like Peter Thiel, or the Andreessen Horowitz fund, in the United States, as well as a whole bunch of libertarian gurus”, notes l French expert, who first quotes Satoshi Nakamoto, with his bitcoin: “He doesn’t trust the traditional banking system, so he creates another one, his own.” Of course, Elon Musk, at the heart of the news with Twitter, SpaceX, Tesla or even Neuralink, fully falls into the category. The subject of the after us titillates, necessarily. “I have the feeling that the fight for the preservation of the environment will accompany us in the next twenty to fifty years. If I had to make a bet, I would say that more collectivist technologies will emerge”, breathes Gilles Babinet. With artificial intelligence?
“We must accept that to innovate, there are accidents”
“I wouldn’t be surprised if the machine learning [NDLR : l’apprentissage automatique, l’une des briques de l’IA] proves to be very effective for the preservation of our environment, he explains. An agriculture that is both productivist and responsible, for example, requires taking into account several thousand parameters, over different time periods, which are therefore very difficult for a human to manage. The machine can. We find the same kind of problem in the management of supply chains, construction, transport…”
The time is, however, for mistrust. The first artificial intelligences are distinguished today by their capacity to be able to write everything and its opposite. Or to generate false images more real than life. The European Parliament is already examining a draft regulation, the AI Act, governing their use and marketing. Which has the gift of annoying Gilles Babinet. “The text is quite good, but there is a risk that it is also good, it is that on arrival, everyone is doing artificial intelligence outside of Europe. In my opinion , we have to accept that in order to innovate, there are accidents. And since we have an older European population, a culture that is probably a little more conservative, we do not easily accept this. innovation, this may seem appropriate to some. I still find that it is an act of defiance towards the future.”
The author suggests waiting. Why not six months to start. “Because generally, after a text is voted on, the legislator does not touch it for several years”. The danger of AI, however, is not a fantasy. “It could be used to easily develop powerful cyber weapons.” Still, in the eyes of Gilles Babinet, it is first of all an extraordinary step forward. “We are already, in France, at the forefront of research, thanks to Polytechnique, Normale sup, Centrale… The question that we must now ask ourselves is the following: how to transform this historic opportunity into productivity gains for our large companies, SMEs and VSEs?I am in favor of a sort of Fourastié mission, named after the economist Jean Fourastié, Planning Commissioner at the end of the Second World War: a government initiative allowing the use of artificial intelligence in the economy and thus gain growth.”
“Age control on the web is a gas plant”
What place does France occupy in the digital world? Or to use a fashionable term: does it show sovereignty? “Beyond the technical considerations, I think that the first form of sovereignty is expertise, so ensuring that civil society largely takes over these technologies. China, for example, has done so when it comes to AI.” The priority of the French government has focused more in recent weeks on a new digital law, in particular establishing age control on pornographic sites. “It starts from real concerns, which are also global. But this technology is a gas plant. For lack of having created a real digital identity beforehand, we put the cart before the horse”, stings Babinet.
The entrepreneur dwells more on France’s ultimately very recent appetite for tech. “It all started, we can say, with the creation of the Public Investment Bank, the BPI, under François Hollande.” This made it possible to produce around thirty unicorns on the territory. “Nevertheless, ours go all over the place, without a specific area of specialization”, while Germany stands out in industry and the United Kingdom in fintech. In short, everything remains to be built. AI, again, should inspire France. A new start-up, Mistral AI, which brings together leading researchers and entrepreneurs, will soon raise 100 million euros to build a large language model (LLM) competing with that of OpenAI (ChatGPT). The very example of what should be encouraged, according to Babinet: real risk-taking, in the right timing, that of the boom in generative AI, and well funded moreover. The next step ? “Making popular the investment in theequity – the actions. Americans do it five times more than Europeans.” And with a bit of luck, “French people may find themselves, tomorrow, shareholders in a company that will be worth tens of billions of euros”.