This is a line of reasoning about the name we give technologies, and about how heavily that name weighs on a country’s future. I’m talking mostly about artificial intelligence, because that’s where the stakes are highest today, and the prompt comes from an episode of Altri Orienti, Simone Pieranni’s podcast on Asia, which sets out from a question that’s specialized only in appearance, what China means when it says artificial intelligence. Altri Orienti, Ep. 176
The Materialist Objection
The first objection to a line of reasoning like this is also the most serious, and it’s worth keeping in hand for everything that follows. Words don’t build data centers, they don’t etch silicon wafers, they don’t train models, they don’t write an industrial policy. A country without chips, without energy, without capital and without infrastructure can choose the most beautiful words in the world and still be left watching. Talk long enough about language and imaginaries and you risk doing exactly what Europe does best, dressing up a material weakness in refined categories: the others build and we interpret, the others invest and we draw distinctions.
The objection holds a truth worth not dismantling, because no technological civilization stands on vocabulary alone. The interesting part begins a step further on. No technological civilization stands without vocabulary either. Words don’t replace matter, they decide how a society arranges matter inside a vision. They decide whether a technology will be felt as a threat or as a prosthesis, as an inevitable destiny or as a commodity like any other. And they decide, before that, who will have the right to speak about it and who will have to obey it.
Every Civilization Has Its Own Technique
The episode I start from serves precisely to show this. It argues that there is no single idea of technology, neutral and valid for everyone, but that every civilization carries its own, born of its history and its image of the world. That’s the sense of the word Pieranni uses, cosmotechnics: the way a culture uses technique depends on how that culture sees the world and the order of things. Said about China it sounds exotic, but it holds for us too. Our own idea of technology, which seems to us the only one possible, is a particular position that learned to present itself as universal.
The first place this difference shows up is translation. The Chinese term 人工智能, which we render as artificial intelligence, doesn’t line up perfectly with our expression, and around that misalignment turns a concept, Machine Decision is Not Final, that uses it precisely to reopen what we mean by artificiality and by intelligence, outside the clichés that dominate the conversation about AI. Urbanomic
Musk and Jack Ma, Two Grammars
The scene that holds it all together is from 2019, in Shanghai, at the World Artificial Intelligence Conference, when Elon Musk and Jack Ma found themselves on the same stage debating the future. They don’t look like two entrepreneurs with different temperaments, they look like two opposite ways of thinking the machine that for a few minutes take human form. Musk talks about superintelligence and danger, about a human life to be carried beyond Earth, and goes so far as to say we might be little more than the biological boot loader for a digital intelligence destined to surpass us. Jack Ma answers with an entirely earthbound optimism, AI isn’t a threat, it’s a tool that came out of human hands, and it should serve to understand ourselves better and to live better here. WIRED reported that meeting as the dialogue between a far more confident Ma and a Musk convinced the machine would outstrip in intelligence even the most intelligent human being. WIRED
It would be easy to dispatch them with a caricature, Musk the Silicon Valley prophet of catastrophe and Jack Ma the optimist of Chinese capitalism. It’s a comfortable reading and it loses the thing that matters, namely that their words move inside two different grammars. Musk speaks as if artificial intelligence were a creature able to detach itself from us and look down on us, already faster and sharper. Jack Ma speaks as if the machine stayed within the human horizon anyway, something we invented and can therefore live alongside, and can hold to account.
The Artificial Kind Isn’t Intelligence
On the same knot, in Italy, the Accademia della Crusca stepped in with an article whose title says it all, Il nome (improprio) della cosa: quella artificiale non è intelligenza (the improper name of the thing: the artificial kind isn’t intelligence). It isn’t nostalgia for the language of the old days and it isn’t the usual war on anglicisms. It’s a more uncomfortable thesis: a wrong name is never innocent. If I call intelligence a system that at bottom computes probabilities, I hand it a mental dignity it may not possess. If I called the same thing, as the article proposes, an artificial simulation of human behavior, I’d look at it right away with less wonder and more caution. The name doesn’t just point to the object, it builds a scene around it, and inside that scene it sets who may enter and on what terms. Accademia della Crusca
Since the Crusca is in play, it’s worth following an etymological trail. The word intelligence comes from the Latin intelligere, which is inter legere, that is to choose among, to grasp one thing by setting it apart from those nearby. At the root, then, intelligence means to sift. It’s the same gesture as the frullone, the flour sieve the Crusca chose as its own emblem to separate the fine flour from the chaff. Calling a machine intelligence, then, isn’t a simple observation, it’s already a choice that carries with it a certain idea of mind, thought reduced to computation and intelligence reduced to a quantity you can measure and grow. Arguing about that name isn’t pedantry. It’s the only way we have to keep a hand on a technology we don’t make at home.
Two Scenes That Pretend to Be Universal
Every name, as I was saying, builds a scene, and the two most powerful scenes today are anything but neutral. In the American one, artificial intelligence is at once a frontier and a risk to the survival of the species, and it turns around the figure of the founder who sees before everyone else. It’s an effective story because it joins a kind of secular religion to a market, with salvation or the end of the world always around the corner. It isn’t a false story, it’s a local story that had the strength to pass itself off as universal. In the Chinese scene, at least as it emerges from this debate, AI is more readily an instrument of government and a proof of the state’s technological strength, inside a planning that as early as 2017 set down a national plan with targets fixed at 2030. DigiChina, Stanford This one isn’t an eternal essence of China either, it’s the product of a precise industrial history and a precise geopolitical competition.
This is where the philosopher Yuk Hui comes back in handy. His idea of cosmotechnics serves precisely to dismantle modernity’s most entrenched conviction, the one that holds there’s a single technique, neutral and universal, identical everywhere and with the same meaning. Even Western technique, Hui says, is a local vision, born of a certain way of separating humanity from nature and of thinking of the world as a reserve to exploit. The paradox is that just as we notice this plurality, the concrete development of AI goes the opposite way, concentrated in a few companies and a few places, with very few languages and very few imaginaries represented. ICI Berlin
Italy Imports, Europe Regulates
It’s from this point that it pays to look at Italy. The Italian problem isn’t only that we invest little, or that we have little industrial scale and little patience. The problem is that we often don’t produce the words we’d need to think our own technology. We import them. We take them from Silicon Valley, pass them through the language of public tenders, flatten them into ministry slides, and then hand them out across the public system like passwords for access to the future. Digital, innovation, transition, platform, ecosystem: words that sound operational and that, repeated often enough, no longer steer almost anything.
And yet a country is shaped this way too, in the everyday vocabulary of the people who administer it and the people who teach it. If a public administration treats artificial intelligence as a magic wand for cutting times and costs, it will build a future made of shortcuts. If instead it treats it as an infrastructure to understand and to govern, it will build a more uncomfortable but more solid one. Almost always, though, it doesn’t treat it at all, it names it only to tick the box on a funding line, and then the future that comes out is just an administrative simulation.
Europe, for its part, has chosen to give artificial intelligence a name mostly through law. The AI Act, the European regulation on AI, has been in force since August 1, 2024 and applies in stages, with different deadlines for banned practices, for the duty to train those who use these systems, for general-purpose models, and for high-risk systems. European Commission It’s no small choice. It’s the attempt to say that technology isn’t only power but also responsibility, and that it has to stay compatible with people’s rights and with real human oversight. It’s perhaps one of the very few real geopolitical positions Europe manages to express on technique.
Here too, though, there’s a risk, and it’s the mirror image of the American one. If Europe speaks of AI only in the language of compliance, it will end up shrinking its own imagination to a procedure. It will regulate beautifully technologies born elsewhere, trained elsewhere, told elsewhere, and sold elsewhere, and so it will do something necessary and insufficient at the same time. Because a civilization can’t limit itself to setting which forms to fill in before using the future others invented, it also has to decide which future it wants to make thinkable.
Linguistic Sovereignty
It’s at this point that the Crusca stops being a detail for academics and becomes a political fact. In a serious country, a discussion about the name of artificial intelligence wouldn’t be dismissed as pedantry, it would enter universities and schools, professional bodies and the firms that write software for others. Not to impose a mandatory word from above, but to force the country into the exercise it has been avoiding for too long, to think before adopting. Because adopting without naming means submitting, it means taking into your home a technology already wrapped in someone else’s vision of the world. When we choose to say AI instead of IA we aren’t only choosing a more international acronym, we’re signaling which center of the world we grant the authority.
None of this is an autarkic defense of the language, which would be ridiculous. Linguistic sovereignty isn’t the rejection of foreign words, it’s the ability not to import, along with the words and without noticing, the whole world they carry with them. A living language takes other people’s words and reinvents them, a subordinate language merely repeats them. And a country that repeats another’s words long enough, sooner or later, ends up repeating its priorities too.
A country’s future is decided at the point where infrastructure and imaginary meet. The data centers are needed, but so is knowing why we want them and which dependencies we’re prepared to accept in exchange. And above all schools are needed, because the way a kid learns the word intelligence today will decide the way they accept, or refuse, the authority of a machine tomorrow.
A Civic Vocabulary of the Future
The real question, then, isn’t choosing between Elon Musk and Jack Ma. Musk reminds us that technique can slip past the ordinary measure of the human, and that not everything we’re able to build stays automatically under control. Jack Ma reminds us that fear can turn into a superstition that freezes everything, and that the machine stays inside a history made of human needs and human work. They each see something and at the same time distort something, each from their own vantage, and the trouble starts when a single vantage claims to speak for everyone.
Italy should then do something at once more humble and more ambitious, build itself its own technical vocabulary of the future. Not a glossary for experts only, but a civic lexicon, in which we go back to asking what we really mean when we say automation, decision, responsibility, error, human oversight, data, delegation. The same doctor, the same judge, the same municipal clerk and the same company that develops software for others will use the same technologies, and that doesn’t mean they have to live them through the same imaginary. I say this from inside the craft. I work as a compliance architect in a small software company in Abruzzo, I build tools that serve to know exactly which components a system is made of, and I spend my days, on platforms for public healthcare and in the legal sector, deciding what things are called, because it’s the name that sets the architecture and the responsibility you take on, and so in the end what will actually get built.
This is why compliance, taken seriously, isn’t bureaucracy. It’s an applied form of humanism, the attempt not to let power come apart from meaning. It’s the moment a society tells a technology it may enter, but not without a name, and that it may also run, but not without answering for the world it’s building.
Maybe the most important question isn’t what artificial intelligence will do to Italy. That question already puts us in the wrong position, as if the future were a storm rolling in from a distant server farm, to be endured and nothing more. The serious question is another, with which words Italy wants to make its own technical transformation governable and open to debate. Because a country that gives no name to its own future doesn’t thereby avoid it, it receives it already named by others, with a hierarchy already decided inside the word, who invents and who merely integrates, who defines and who merely translates.
The future begins before the technology. It begins the moment a community finds, or loses, the words to say what it is becoming.
Key takeaways
Words don’t build data centers, but they decide how a society arranges matter inside a vision. They set whether a technology will be felt as a threat or a prosthesis, and before that, who has the right to speak about it and who must simply obey it.
There’s no single idea of technique, neutral and valid for everyone. This is Yuk Hui’s cosmotechnics: even our own technology, which looks to us like the only one possible, is a local position that learned to present itself as universal.
A wrong name is never innocent. Calling intelligence a system that at bottom computes probabilities hands it a mental dignity it may not possess. Arguing about that name isn’t pedantry, it’s the only way we have to keep a hand on a technology we don’t make at home.
Italy often doesn’t produce the words it would need to think its own technology: it imports them from Silicon Valley, flattens them into the language of public tenders, and hands them out like passwords for access to the future. A country that repeats another’s words long enough ends up repeating its priorities too.
Europe has chosen to name AI through law, and the AI Act is perhaps the only real geopolitical position it manages to express on technique. But if it speaks of AI only in the language of compliance, it will shrink its own imagination to a procedure: it will regulate beautifully a future invented elsewhere.
Questions & answers
What is cosmotechnics?
It’s a concept from the philosopher Yuk Hui, taken up in Simone Pieranni’s podcast Altri Orienti. It names the idea that there is no single neutral and universal technique, identical everywhere, but that every culture uses technique starting from how it sees the world and the order of things. Even Western technique, Hui says, is a local vision, born of a certain way of separating humanity from nature and of thinking of the world as a reserve to exploit. The uncomfortable consequence is that our idea of technology, which seems to us the only one possible, is just a particular position that learned to present itself as universal.
Why does the Accademia della Crusca say the artificial kind isn't intelligence?
In an article titled Il nome (improprio) della cosa: quella artificiale non è intelligenza (the improper name of the thing: the artificial kind isn’t intelligence), the Crusca argues a thesis more uncomfortable than the usual war on anglicisms: a wrong name is never innocent. Calling intelligence a system that at bottom computes probabilities hands it a mental dignity it may not possess. The word intelligence comes from the Latin intelligere, that is inter legere, to choose among, to sift: the same gesture as the frullone, the flour sieve the Crusca chose as its emblem. Calling a machine intelligence is already a choice, it carries with it thought reduced to computation and intelligence reduced to a quantity you can measure and grow.
What does China mean when it says artificial intelligence?
The Chinese term 人工智能, which we render as artificial intelligence, doesn’t line up perfectly with our expression, and around that misalignment turns a reflection that reopens what we mean by artificiality and by intelligence. In the Chinese scene, at least as it emerges from the debate this essay starts from, AI is more readily an instrument of government and a proof of the state’s technological strength, inside a planning that as early as 2017 set down a national plan with targets fixed at 2030. It isn’t an eternal essence of China, it’s the product of a precise industrial history and a precise geopolitical competition.
What is linguistic sovereignty?
It isn’t an autarkic defense of the language, which would be ridiculous, and it isn’t the rejection of foreign words. It’s the ability not to import, along with the words and without noticing, the whole world they carry with them. A living language takes other people’s words and reinvents them, a subordinate language merely repeats them. And a country that repeats another’s words long enough, sooner or later, ends up repeating its priorities too.
Is the AI Act enough to give Europe a position on technology?
The AI Act, in force since August 1, 2024 and applied in stages, is perhaps one of the very few real geopolitical positions Europe manages to express on technique: it says that technology isn’t only power but also responsibility. But there’s a risk, the mirror image of the American one. If Europe speaks of AI only in the language of compliance, it will end up shrinking its own imagination to a procedure, and it will regulate beautifully technologies born elsewhere, trained elsewhere, and told elsewhere. A civilization can’t limit itself to setting which forms to fill in before using the future others invented, it also has to decide which future it wants to make thinkable.
The author
