'Ai, train the man or productivity will not increase'

However, when an investment in Ai costs up to one trillion, it is no longer just a question of ethics but also of industrial, political and cultural choices. And it is no coincidence that Giuliano Noci, pro-rector of the Milan Polytechnic, has called for overcoming the false dilemma between man and machine: 'There is no real dichotomy. Artificial intelligence must be a co-pilot at the service of man'. Mr. Noci spoke of the 'digital transition' as a phenomenon that can undermine Western cultural identity to the extent that it is a transition that, in other regions of the world, is approached with less resistance: 'In Asia, for example, a cultural model prevails that exalts connection and does not put the individual at the centre, but his social projection'.

So what can happen and with what risks? Massimo Lapucci, co-author with Stefano Lucchini of the book 'Rediscovering the Human', approached the topic from the perspective of finance, where the use of artificial intelligence is already widespread. He put it bluntly: 'We cannot have blind faith in algorithms. Ai must be guided and monitored'. And if 'in areas such as credit scoring or portfolio management AI has become central, the models must be understandable and monitorable. You cannot delegate everything to the machine without knowing how it works'.

The logical thread thus leads the economist to draw attention to governance and the risk of systemic bias: 'Artificial intelligence is a tool that can discriminate if not properly designed. It is crucial to have rules, such as those proposed by the European Ai Act'. However, this emphasis is also accompanied by a proposal: to place curators alongside Ai evaluators, figures capable of ensuring that the machines are truly at the service of man.

The latter remains at the centre and must be aware, Father Benanti points out, of a risk that cannot now be dismissed: 'In the Middle Ages illiteracy was not knowing how to read manuscripts,' he says. 'Today it is not knowing how to interpret digital information. The new generations risk information illiteracy if they are not educated to distinguish the value and credibility of sources. It is an educational problem, even before the technological one'.

Ultimately, however, Noci concludes, 'there is no need to focus on the theme in terms of the antinomy between man and machine. The theme is to find a Hegelian synthesis between man and machine. Technology can indeed become a form of order if properly designed and governed'. And who knows to what extent the Ai Act will be able to do the latter, about which the Vice-Chancellor of the Milan Polytechnic, however, is not soft: 'The Ai Act defined by Europe is a form of regulation in which those who regulate do not control the playing field. So European companies enter crippled. Elsewhere in the world, they enter the field with doping'.