OPINION. AI: “Like a little music”

By Philippe Boyer, director of institutional relations and innovation at Covivio.

Two recent news stories show that AI, once adored because it brings certain progress to humanity, no longer benefits from the same appeal. The first, dated April 10, when a young man threw a Molotov cocktail at the gate of Sam Altman’s residence in San Francisco, before threatening the OpenAI headquarters. No one was injured but the symbol struck home: the boss of the company to which we owe ChatGPT thus became the physical target of anger directed against AI and its figureheads, a few weeks after OpenAI had signed an agreement with the American Department of Defense.

Another scene, another setting, a month later at the University of Arizona. Coming to attend a graduation ceremony, Eric Schmidt, former boss of Google, was booed when he presented AI as a major transformation in the future of students. Here again, the symbol is strong. Young people, who were believed to be spontaneously won over to digital innovations, no longer necessarily applaud the prophets of Silicon Valley. She interrupts them. She challenges them. She answers them that the promise of an “augmented” future sometimes resembles the announcement of a confiscated future.

These two episodes say something profound. Until recently, criticizing artificial intelligence exposed one to being seen as nostalgic, a technophobe, a gloomy mind clinging to the old world. The dominant narrative was clear: AI is coming, it is inevitable, we must adapt because it will transform everything. Any reservation became suspect. All caution looked like defeat. Any questions about its social, cognitive or democratic effects were placed in the museum of human fears worthy of the Luddite movement, these machine-breaking workers in England at the beginning of the 19th century.

However, recently, something has changed. A little music of refusal begins to be heard, via several channels. Certainly, it is not yet a great choir, but that does not prevent it from becoming audible. It crosses campuses, businesses, political assemblies, “white-collar workers”, artists, teachers, young graduates. It sometimes expresses itself in a confused, even violent manner, but this “little music “technocritique” says something profound: AI must not become the polite name for our replacement.

The most striking thing is that this concern comes primarily from young people. A generation that had adopted video games, social networks, platforms and smartphones does not look at AI with the same enthusiasm. For what reason? Perhaps because this technology does not only promise new uses; it also threatens his future integration into the professional world.

It is now obvious that we can read in the press or in numerous studies on the impact of AI in the workplace: the tasks entrusted yesterday to juniors – writing, synthesizing, coding, analyzing, producing a note, preparing a file – are also those that AI automates the fastest. In recruitment, the situation becomes absurd: candidates use AI to write their letters; employers use AI to analyze them. In short, the machines respond to each other, while humans fade behind these machines.

This distrust of AI is also fueled by a paradox: the tech giants have blown the flames by multiplying declarations, sometimes grandiose or apocalyptic, depending on the moment. Elon Musk predicted that almost all jobs could disappear. Sam Altman compared AI to near-nuclear power. Dario Amodei, boss of Anthropic, spoke of the dangers of overly powerful models, particularly if they were used in weapons or cyber security systems. In short, and by dint of presenting AI as a force capable of fundamentally transforming the economy, employment, democracy, war… its promoters cannot today be surprised that the general public ends up believing them while questioning the meaning of this future society where obsolescence does not come not so much machines as humans.

The problem is that we cannot sell the apocalypse to investors on the one hand and appeasement to citizens on the other. To attract capital, justify dizzying valuations and prepare for future IPOs, we must convince that AI is a structural revolution comparable to electricity or the Internet. Saying: “We are going to improve a few processes” is not enough to raise hundreds of billions. Conversely, asserting that: “We are going to transform the world” works better. When we repeat that AI will have a structuring effect, we should not be surprised that the world is questioning itself and demanding accountability or at the very least that the potential fruits of this future transformation are known, anticipated and shared.

Finally, another signal, more geopolitical, confirms that AI has changed category. When the White House recently pushed Anthropic to cut off access to its latest models to foreign nationals, including to certain developers working on American soil, the message sent is clear: AI is becoming a power resource, comparable to critical components or military know-how.

The example of this forced dismissal of an AI by political decision ruins the illusion of an open world where everyone has free access to the best tools. AI is no longer just about innovation; it becomes a matter of sovereignty. Those who own the models, the data, the chips, the infrastructure and the capital own a share of the future power. Others will have to ask for access, accept restrictions, and endure dependencies. For Europe, we already knew this but the message is clear: we do not defend our values ​​with technologies that others can take away from you overnight.

But the heart of the subject remains anthropological. What worries us is not that AI does things for us. It’s that it ends up taking our place. Yesterday, the machine replaced muscle. Today, it absorbs language, reasoning, evaluation, creation, sometimes even relationships. It is no longer content to execute; it is no longer content to calculate; it is no longer content to respond; it is before our eyes: assistance must not become a substitution.

AI, and that’s good, can help treat, search, translate, secure, diagnose, model, write, learn. In some areas, not using it is even irresponsible. But it must remain an instrument, not become an authority. It can clarify a decision, not confiscate it. It can increase human capacity, not organize the withdrawal of humans.

This little music of refusal is therefore not necessarily bad news. It can be the beginning of maturity. After technological intoxication, discernment. After the prophecies of Silicon Valley, the return of politics. After the promise of an increased humanity, the time for answers to these simple questions: increased by whom, for what, and to what extent? AI will go too far and will be increasingly challenged if it convinces us that we have become incidental. Progress consists of ensuring that, in the face of increasingly powerful machines, humans remain the subject of their own history. This is the music we want to hear.