News

Artificial Intelligence: A User’s Guide

by

Mario De Caro is one of Italy’s leading scholars in AI ethics. A philosopher and Professor of Moral Philosophy at the University of Rome Tre, where he holds the UNESCO Chair in Ethics of Artificial Intelligence and Practical Wisdom, and a Visiting Professor at Tufts University in the United States, he has long explored the relationship between mind, technology and society. Starting from the origins of the “mechanisation of thought”, he explains how AI has already reshaped our cognitive paradigms, the risks posed by the recent rise of agentic AI, and why collective policies and regulations are needed. When did the idea of mechanising thought first emerge? Philosophers such as Pascal and Leibniz had already envisioned systems capable of carrying out reasoning through mechanical processes. The real turning point came in 1956, at the Dartmouth Conference, where the term “artificial intelligence” was officially coined. The original idea was to represent all human knowledge symbolically and apply logical rules to infer new information. It was a rule-based form of AI, grounded in deductive reasoning. What distinguishes today’s AI from its earliest forms? The difference is profound. Early AI systems followed predefined rules: they started from axioms and produced logical consequences. Today’s systems, by contrast, are data-driven and learn from experience. The decisive shift came in 2017 with the development of new neural network architectures and Large Language Models (LLMs). Rather than operating sequentially, these models process vast amounts of data in parallel, identifying patterns and probabilities. In this sense, AI reasons abductively: it generates the most plausible hypothesis based on the available data. This is a form of reasoning often used in science. Darwin, for example, developed the theory of evolution by connecting phenomena that initially appeared unrelated. Abductive reasoning is powerful because it can generate innovation, but it also carries risks: its conclusions are never certain. Do Large Language Models truly understand language? The most advanced systems display increasingly sophisticated forms of functional understanding. They can generate text, images, music and even creative solutions in ways that are often indistinguishable from those produced by humans. We are no longer dealing with merely advanced calculators. Systems such as AlphaZero have demonstrated remarkable strategic and creative capabilities, learning autonomously through trial and error. Can we speak of conscious artificial intelligence? Scholars remain divided on the issue: some believe it may be possible, while others reject the idea entirely. The crucial point, however, is that AI can be highly powerful—and potentially dangerous—even without consciousness. Today’s systems are inherently indeterministic: we cannot fully predict their behaviour, nor can we always understand how they arrive at a particular decision. Agentic AI goes beyond simply responding to prompts. It can make decisions, plan actions and operate autonomously. This creates enormous opportunities, but also significant risks, as such systems are already being deployed in strategic domains including finance, healthcare, public administration and defence. What are likely to be the next major areas of application? Medicine is likely to be the field where we will see the most significant impact. AI is already accelerating both diagnostics and pharmaceutical research. But its applications extend across virtually every sector: the creative industries, logistics, financial services and education, to name just a few. The challenge is that these systems are often opaque: they work, yet we cannot always explain precisely why they produce certain outcomes. This is the issue of explainability. In a sense, both humans and machines are opaque systems. AI generates solutions through internal processes that are difficult to decipher, just as we do not fully understand the deeper motivations behind our own behaviour, often constructing rational explanations only after the fact. Yet it is precisely from the interaction between these two forms of opacity that some of the most interesting results may emerge. At the same time, significant risks remain, including the erosion of critical autonomy, information manipulation and algorithmic discrimination. AI can amplify dynamics already present in social media, such as polarisation and echo chambers. There is also the problem of “hallucinations” — outputs that are false, yet highly plausible. What impact will AI have on the creative disciplines? Artificial intelligence can generate images, music and texts that are often indistinguishable from those created by humans. There have been instances in which poems produced by ChatGPT were mistaken for the work of renowned authors and were even highly rated by readers. We are not yet talking about the radical creativity of a Shakespeare or a Caravaggio, but rather about a form of “everyday” or “average” creativity that these systems have already demonstrated they can achieve. For this reason, it is essential to study how such systems work in order to develop a critical and genuinely human perspective, grounded in the capacity for judgement—the ability to recognise when something may appear convincing or correct, yet ultimately fails to hold up under closer scrutiny. What will the impact be on the brand identity sector? Where will the value of creative agencies lie? Interaction is the way forward. AI is an extraordinarily powerful tool, but it will not replace creative professionals. The added value of creative agencies will lie in their ability to integrate and harmonise these technologies within organisational processes, acting as a critical interface between artificial systems and the ways they are actually deployed and used. Today, there is growing demand for this kind of strategic guidance. Ultimately, machines can stimulate and enhance creativity, even within corporate decision-making processes. Yet understanding how to make this happen in practice is far from straightforward. What we are facing is an entirely new anthropological and operational paradigm—one that requires not only technological expertise, but also judgement, interpretation and a deep understanding of human behaviour. How can we govern artificial intelligence? Are the current regulations sufficient? No. The European AI Act (Regulation (EU) 2024/1689) is an important first step, but it is not enough. What is needed are international frameworks and, above all, systems designed with intrinsic ethical safeguards—an approach often referred to as “ethics by design”. Most importantly, effective governance … Read more

1946 – 2026: Eighty years young. Happy birthday, Vespa!

by

On April 23, 1946, the patent for the Vespa was filed—the scooter that would go on to become a global symbol of Italian design and lifestyle. The project was commissioned by Enrico Piaggio with the aim of supporting affordable individual mobility in a country devastated by the war. Corradino D’Ascanio, a brilliant aeronautical engineer who had previously developed a helicopter prototype and patented the variable-pitch propeller, approached the challenge by setting aside conventional motorcycle design principles (typically derived from bicycle frames). Instead, he drew on his aeronautical expertise and on surplus components that had become unusable. Until then, in fact, Piaggio had produced military aircraft, ships, and railway carriages. Thus, the starter motor of radial aircraft engines became the propulsion unit of the new scooter. Similarly, the small wheels with a single side-mounted suspension system were derived from aircraft landing gear. The true innovation lay in the frame—a monocoque body structure that gave the Vespa its distinctive shape and introduced entirely new user-oriented solutions. The riding position was designed around a “natural” seated posture and did not require the rider to straddle the fuel tank, as on motorcycles. Mechanical parts were enclosed within the bodywork (as in cars), and the front shield protected the rider’s legs from mud—and even in the event of a fall. In short, it was a revolution: a vehicle designed for ease of use, accessible even to women, as it could be comfortably ridden while wearing a skirt. The tribute to the Vespa, created by Inarea, dates back to 1997, marking the return of the scooter’s historic name after the “Cosa” phase introduced in 1988. It consists of 16 images—compositions created by assembling coherent everyday objects to reconstruct the iconic silhouette. The visual grammar is characteristic of Inarea, where “humble” elements of daily life are combined to generate new forms—in this case, the Vespa myth. These interpretations were met with immediate acclaim, leading Piaggio to turn them into a calendar distributed worldwide, titled “The Vespa 1998 Calendar.” They are still on display at the Museo Piaggio. Happy birthday, Vespa—eighty years young.

Projects  

Subscribe to our newsletter