The session was moderated by Stefano Martinotti, an artificial intelligence researcher and former McKinsey partner, alongside three prominent voices: Mario Margotta, Professor Giuseppe M.J. Barca and Richard Ralphsmith.
“Today, we hear diametrically opposed views about AI,” Martinotti began, alluding to one side arguing that the promise of a new golden age of prosperity is upon us while the other warns a rushed adoption could “tear the very fabric of our society”.
Yet the real issue, he explained, is semantic before it is technological. Artificial intelligence has existed for more than eighty years; what is fuelling the current debate is Generative AI, a far more recent architecture (post-2015) capable of working with unstructured data—texts, forms, images—and generating outputs consistent with its training.
The numbers are striking. Recent estimates place the annual economic potential of GenAI at over $4.4 trillion.
Compared with a global economy projected at $117 trillion by 2025, that would amount to adding the equivalent of “two Italies” to global GDP every year—figures that, for better or worse, demand serious reflection.
This impact unfolds along three main lines: the automation of repetitive tasks, increased human productivity and faster innovation cycles through the extraction and indexing of knowledge.
“Eighty per cent of the workforce will see at least 10 per cent of their tasks supported by AI, 50 per cent of activities will be completed faster and 60 per cent of annual productivity growth will be driven by GenAI and automation,” Martinotti explained.
But not everything is straightforward. Margotta highlighted concerns around reliability, cost and data governance.
“More than 40 per cent of companies have already reported cases of improper use of GenAI,” he noted. Businesses want clarity on where their data goes and how it is used.
While industry seeks certainty, research is opening disruptive possibilities. Professor Barca pointed to the pharmaceutical sector, where bringing a drug to market typically requires 10 to 15 years and around $2 billion in investment.
With computational simulation and GenAI, it is now possible to design and test compounds virtually with accuracy comparable to physical experimentation, potentially reducing development timelines by 50 to 75 per cent.
“It’s not just about doing things faster and more cheaply,” he said. “It’s about opening the door to treatments that five or six years ago were simply unimaginable.”
From the perspective of marketing, Ralphsmith offered a more counterintuitive reflection: as the cost of creative production falls, advertising becomes ubiquitous.
He explained that when technology makes everything accessible, what ultimately differentiates brands is human connection.
Amid these developments, local institutions also have a role to play: ethical regulation, dialogue with academia, internal coordination and the sharing of best practices.
Because the challenge, it is increasingly clear, will require collaboration across multiple fronts.
