Neither apocalypse nor utopia on the future of artificial intelligence and technological progress: this is the approach that the Digital power, the interesting book by the professor Luisa Torchia, a student of Sabino Cassese and full professor of Administrative Law at Roma Tre University, recently published by il Mulino. The essay systematically yet easily readably addresses all the highly topical issues surrounding AI and the opportunities and risks of technological progress. But what exactly is Digital Power? "It is private power, concentrated in the hands of a few American and Chinese companies that have reached such a size that they can dominate markets and compete with states" and beyond. An enormous power that requires effective regulation, precisely what Trump's America doesn't want. But the challenge isn't over. Here's what Torchia argues in the interview with FIRST online.
Professor Torchia, significantly the last chapter of your book “Digital Power” opens on a burning topic of current interest in these days and is entitled The digital future between apocalypse, utopia, and realismIt seems you lean toward the third option, but how do you actually see the future of digital power?
It is a future yet to be written and depends significantly on several factors. It depends first and foremost on the willingness of societies, states, international organizations, scholarly communities, and individuals not to be, so to speak, "disarmed" by the apocalyptic prospect, which, as in all apocalypses, produces great anxiety and inhibits all action, because the apocalyptic prediction assumes not only that the world is about to end, but also that this end is inevitable. And, as Christian Friedrich Hebbel astutely observed almost two centuries ago, "There are people who would even be consoled by the end of the world, as long as they had predicted it.“The apocalyptic perspective is therefore also, paradoxically, consolatory, because it justifies the inability or unwillingness to act and react.”
So what is needed?
At the same time, we must not be fooled by promises of great and extraordinary future benefits for all. The utopian vision of technological development is belied by the many unfulfilled promises. No new life-saving drugs or treatments for cancer and degenerative diseases have been discovered, no solution has been found for climate change, no low-cost energy sources have been identified, no new art, or new literature, or new music of comparable quality to human production has been "generated." And one might wonder whether the declarations of big tech companies about the "magnificent and progressive fortunes" of technological development—always set several years or even decades from now—aren't simply advertising messages aimed at attracting the capital needed for enormous investments whose returns are also postponed into an indeterminate future.
What do you think?
I believe that a realistic approach is not only possible, but also necessary. It doesn't deny the enormous advantages and benefits of technological progress, but focuses first on the real advantages, not the promises. It then takes into account both the risks new technologies bring, as well as the completely unequal distribution of those advantages and risks. There are groups and companies that largely capitalize on those benefits and shift all the risks and disadvantages onto other groups and communities. The perception of this phenomenon is widespread today and explains the growing distrust of large tech companies, the opposition to the construction of data centers, now very strong even in the United States, and the proliferation of lawsuits against platforms for violations of competition law, intellectual property rights, privacy, and for harming the mental health of young people.
A difficult road to travel?
It's obviously a difficult path to follow, but big tech companies themselves are beginning to feel the pressure of an increasingly hostile public opinion and are clearly seeing the risk of ever-increasing—and costly—liability that could be attributed to them. Paradoxically, but not entirely paradoxically, they see the solution in public intervention, both regulatory and financial, increasingly advocated by those who have so far fiercely opposed attempts at external regulation, emphatically asserting that self-regulation was sufficient. State-imposed rules can also be misused to mitigate competition among large players and disadvantage smaller ones, simultaneously raising barriers to entry for new entrants. Rules must therefore be very carefully and accurately established and subject to review to ensure their effectiveness and allow for rapid correction.
What are the dominant features of digital power?
“Synthesizing the results of the analysis carried out in my book, digital power can be defined as a private, global, platform-structured, extractive, normative, conformative and conformist, persuasive and pervasive power.
It is, first and foremost, a private power, concentrated in the hands of a few American and Chinese companies, which have reached such a size as to allow them to dominate the markets and compete with the States."
Impressive.
It is, then, a power with a global scope, operating without territorial borders and offering or selling products and services that have become essential to individual and collective life throughout the world. It is a power structured according to the platform model, as a meeting place for supply and demand, for information, goods, and services, which has become so indispensable for both commercial and end users that it prevents or at least makes it very difficult to exit the dominant platform, partly due to the lack of alternatives. Systems based on LLMs are also integrated into existing platforms (like Gemini for Google) or pursue the platform model, offering an increasingly wider range of integrated services that cannot be accessed separately. It is an 'extractive' power, which seeks to obtain the resources necessary for its exercise without bearing the costs and by appropriating the revenues to an extreme extent. It is a normative power, because it establishes the rules for interactions on platforms and, with the most recent applications, also rules for how medical, banking, or financial activities are practiced. It is a conformist and tendentially conformist power, because through models it reduces the variety of reality and then returns a simplified and uniform version of it, based on a merely statistical and not moral, ethical, or social order. It is a persuasive and pervasive power on the economic, political, and social levels, because the technology on which it is based has now become an integral part of everyday life, with the growing use not only of specialized applications, but perhaps even more so of generative AI capable of interacting through natural language.
In your book, you illustrate very clearly how digital power has become political power and how digital capitalism has become more powerful than states and is endangering democracy itself: is this a challenge we have lost forever, or can we still hope to rebalance the balance of power between digital power and democratic power?
A distinctive and characteristic element of digital power is concentration: a concentration so strong and so impenetrable that neither the market nor the state have been able, at least for now, to produce effective measures of redistribution. Concentrated power dangerously resembles unlimited power, and unlimited power is almost by definition arbitrary power. Like all powers, however, digital power also requires legitimacy, and this legitimacy today cannot be based solely on force: technology companies also need to sell products and services on the market, to convince consumers that their products and services are useful and indeed essential, just as they need to raise capital and deliver on promises of returns, to convince communities to accommodate their data centers, to convince states that they are not endangering their sovereignty. It is no coincidence that the 'bosses' of big tech often fantasize about new planets or areas of this planet reserved for them.
Why do they do it?
Because they perceive society and its rules as a constraint they don't want to submit to, one they would like to free themselves from. This, however, is a completely illusory perspective, as demonstrated by two different reactions. In autocratic countries, such as China, the state exercises firm control over technology companies and the use and diffusion of technology. I certainly don't want to point to autocracy as an example, but it demonstrates that public power, if there's the will, is not impotent. Indeed, in the case of China, the state requires that artificial intelligence models reflect the 'values' of socialism, demonstrating that technology is not neutral. In democratic countries, there is growing intolerance and distrust towards technological innovation that destroys jobs, creates inequality and discrimination, accentuates poverty and the marginalization of entire groups and communities. Democracy, with its inclusive, pluralistic, and deliberative capacity, is a possible antidote and could be strengthened precisely as a tool that allows modern societies to defend themselves and preserve values. and collective goals. Hope, therefore, is possible, but it must be cultivated, exercised, and defended, and this is a task that falls to each of us, as citizens of democratic countries."
To govern digital power without stifling technological innovation, yet without being overwhelmed, you argue that regulation is the key. In this area, Europe is ahead, while Trump's US is reluctant to do so. But if the regulation of digital power can't extend to the domestic market, it seems inevitable that its effectiveness will be limited. Or not? How can this misalignment between Europe and the US be overcome?
"First of all, it could be a contingent misalignment: it depends a lot on what happens in the American elections starting in November and in the US over the next two years. Regulating innovation is difficult because the subject matter of the rules is constantly changing, and for this very reason, the European regulatory approach to artificial intelligence is essentially risk-based: risks cannot be avoided, but they can be mitigated and, above all, better understood if there is a control system that detects them. European regulation can be criticized for its burdensomeness—and an attempt at simplification is being made with the Digital Omnibus—but it has an important advantage: no big tech company can give up the European market, one of the largest and richest in the world. And so, despite much criticism and resistance, no big tech company has left Europe, despite the regulation. The misalignment between the US and the EU, therefore, exists but does not prevent the application of European rules, within Europe and sometimes even beyond its borders, because many legal systems are converging around the same type of principles and tools for regulating and controlling artificial intelligence. Also The argument that regulation hinders innovation becomes less convincing when innovation emerges from its embryonic stage, as has now happened with artificial intelligence. Regulation is the main tool through which an industry can guarantee the safety of its products and gain consumer trust: no one would board a plane or buy medicine unless they were reasonably certain that the journey and treatment were safe and risk-free.
What do you think is the best option?
Naturally, the best option would be a global regulation, joined by many states and international organizations. This convergence is not impossible. In my book, I give two examples where this convergence has occurred: atomic energy and cloning. The differences between these two examples are clear: atomic power tends to belong to states, while the cloning capacity arises in research laboratories. In the first case, states have resorted to international treaties, while in the second case, there has been a convergence on the choice, adopted individually by each state, to ban reproductive cloning. In both cases, however, the perception of the dangers and risks involved—in the case of the atomic bomb, dramatically demonstrated in Hiroshima and Nagasaki—together with the belief that isolated national solutions would be completely unsuitable—has led to the adoption of shared rules that states and researchers have accepted and which, at least for now, have not been violated or, in the case of possible violations, as in Iran with regard to atomic energy, have been subjected to enforcement mechanisms. international controls. In both cases, to put it very simply, these are rules fundamentally based on prohibitions and the acceptance by the parties involved of the exercise of a power to monitor compliance with the prohibitions by national or international authorities or agencies.
But has an attempt at convergence on AI been made?
The G7 also attempted to establish international guiding principles and shared rules of conduct regarding technological innovation in 2023, adopting principles appropriately named the Hiroshima AI Process. These international guiding principles should apply to the entire AI lifecycle (and therefore to all stakeholders involved in the various phases), and the International Code of Conduct provides guidelines that AI developers can voluntarily adhere to. The implementation of the Hiroshima AI Process is entrusted to the OECD, which defined the framework for monitoring reports in February 2025.
The challenge to digital power is being played out, as your book also shows, on an international level, but what can Italy do within Europe to put itself in the best position to play the game?
Italy doesn't have large amounts of investment capital or natural resources, but it has excellent universities, and is capable of training excellent researchers—who go abroad to seek their fortune: we train them and then pass this heritage on to other countries. It has a tradition of culture and technical expertise that, despite the destruction of so many industrial experiences in recent years, still endures. If, instead of dispersing this wealth of culture and expertise, it were concentrated in research and production centers, allowed to grow to a critical mass, and networked with the manufacturing system and the public sector, we could make a significant contribution.
