Artificial Intelligence: is France up to the task?

While Elon Musk announces, with much media buzz, the launch of his company NeuraLink, intended to develop an implant that would allow our brain to communicate with an Artificial Intelligence, the French government is thinking. It has commissioned a report from 500 experts, researchers and representatives of the business world on the subject. "France IA" has just been published and it leaves us hungry.

 

À At the request of Thierry Mandon, Secretary of State for Higher Education and Research, and the former Secretary of State for Digital Technology Axelle Lemaire (since replaced by Christophe Sirugue), more than 500 experts, researchers, business leaders and members of parliament worked on issues related to Artificial Intelligence (AI). One question was asked: how to develop research and innovation in France in the field of AI?

In response, the report defined a very ambitious field of research, ranging from the establishment of a French mapping of AI to an inventory of research projects, through the modalities of industrial applications or the implementation of a public debate on the subject. All these issues are necessary to draw up the future national AI strategy. A strategy based on 59 recommendations.

 

This may be the latest initiative of the current government. Is it not coming a little late? The President of the Republic, François Hollande, has made AI a national issue in which our country has many assets. But he admits that the risk would be " be content with the status of a consumer captive to innovations conceived and developed elsewhere ». It's just time to make our country what the president dreams it is. " a key producer who will help shape the future face of artificial intelligence, and thus the world to come.. »

 

The problem with AI is that technical progress is inevitably accompanied by ethical, social and societal questions. This is a good thing, but not all countries in the world are on the same wavelength in this respect. Enthusiasm for technological prowess is loudly proclaimed, but doubts and fears, if not more or less rational fears, always arise. So, the France IA report seeks to clarify the current debate on artificial intelligence by highlighting the challenges and future potential that mastery of these technologies represents. It also wants to mobilize the AI community in France and internationally and define a plan of action and mobilization of funding, in particular through the Future Investment Program. In short, the report aims to assert France's leadership in Europe and the world.

 

These (pious?) wishes are based on a number of realities. Yes, France has a very dense network of more than 250 research teams working on pure AI, not to mention the thousands of researchers in the human and social sciences working on the subject. It is also true to say that France is present in almost every sector impacted by AI, whether it be robotics, human-machine interaction or AI law. 270 French startups are actively working on these issues. But, if it is well argued, documented and well structured, this report sometimes lacks lucidity.

 

Indeed, isn't this French initiative already too late? As far as AI is concerned, the preponderant part of research is carried out in the private sphere, dominated today by Americans and certainly tomorrow by Chinese companies. Groups such as Facebook, Microsoft, IBM, Google to name only the best known, but also the Chinese Baidu or Xiaomi inject billions of dollars into AI projects. Money isn't everything; they poach the best researchers all over the world. Last year, Facebook opened its dedicated AI center in Paris, hosting several dozen of the best French specialists. At the head of the Facebook AI team, a Frenchman: Yann LeCun.

Within the framework of the Parliamentary Office for the Evaluation of Scientific and Technological Choices (OPECST), the French senators have also published an reportthe day after the government's presentation. As soon as their investigation began, they denounced: " French laboratories are plundered of their researchers by North American and Chinese multinationals.. "And this billion-dollar research is making tremendous progress. There's been a lot of talk about that computer that flatly beat the world Go champion. But every day, new advances are announced in every field: in robotics, in vehicles, in health care, etc. All areas of society are impacted by AI algorithms.

Admittedly, one can put a damper on these thunderous announcements. This is what Jean-Gabriel Ganascia, an expert in this sector, often tries to do. He tries to reassure by explaining in his latest book The myth of the SingularityThe moment of the Singularity, the moment when the machine's intelligence will surpass that of humans, is not for tomorrow morning. The fact remains that progress is being made, without France.

 

But the real danger - immediate, this one - of AI does not come from the fact that it could replace humans, but from the superpower that the large multinationals in the sector are acquiring. These companies have acquired a size and power that rivals that of the States. Moreover, their ambition is to replace the large States. In an interview with the weekly newspaper l'Express, Jean-Gabriel Ganascia explains: " For example, by offering (paid) services that relate to the major sovereign functions, such as security with facial recognition. ». He goes on to say: « States are unable [to produce these services themselves], limited by the lack of a database and laws that protect citizens in principle. They [Facebook, Google] don't have this problem, because we are the ones who give them the photos! "

 

This question of state sovereignty undermined by AI deserves answers to match. All the more so as the impacts on society are beginning to be felt everywhere, with job losses due to robotization in particular.

Are the responses provided by the public authorities, if necessary, sufficient? To feed into the government report France IAFrance Stratégie has also split by one report on the subject. Co-authored by startuper Rand Hindi of the Conseil national du numérique (CNUM) and Lionel Janin of France Stratégie, the report points out the measures to be implemented to "improve the quality of the digital environment". anticipate the economic and social impacts of AI ». A title that says a lot about what the authors think of a fight lost in advance. Certainly, the report emphasizes the importance of the data that make up the AI nutrient. It proposes to ensure that " the machine remains at the service of man ». He finally preaches for " a broad consultation on artificial intelligence and the changes in employment resulting from it ".

By way of remedy, the emphasis is placed on formation, always formation, always formation: " Because tasks will continue to change, we need more than ever individual lifelong learning. If we fail to anticipate, we could find ourselves faced with job losses; if we fail to mobilise, we run the risk of economic and strategic dropout.. »

A diagnosis and solutions that reveal the extent to which the emergence of AI and its dazzling progress leave our rulers dumbfounded if not disarmed, leaving them with no other option than to try to contain as best they can an irresistible and above all inevitable surge.

 

What to do in the face of this fight that seems to be lost in advance? Elon Musk, the billionaire with prolific ideas, boss of Tesla, Space X, the one who wants to send colonies of Earthlings to Mars, has just proposed an idea that has caused a lot of noise: to equip us with electronic implants in the brain to increase our intelligence.

The announcement of the launch of his company NeuraLink has filled with satisfaction all supporters of transhumanism. It is astonishing, coming from a man who has always declared himself suspicious of AI, even sensing, like Stephen Hawkins or Bill Gates, that it was the greatest danger threatening humanity. So why this apparent reversal? Perhaps because Musk knows that the fight to contain the AI's advance is already lost. And that the only way we can save humanity from the future peril of a strong AI is to be smarter than machines. Why not using microchips and algorithms to boost our neurons?