The production system at the challenge of digital transition

One figure, however, is decisive for understanding the scale of what is happening. As early as 2030, it is estimated that 25 per cent of the gross domestic product will be determined by the further development of this technology. In this, it is clear that Italy on the foundations of data calculation models is certainly lagging behind, but, explains Nicola Gatti of the Politecnico di Milano, 'this does not mean that there is not great scope for making the models we use adaptable in an almost tailor-made manner, while at the same time needing to be interpreted and dropped into different realities'.

According to Chiriatti, this is a great opportunity for the made-in-Italy industry to implement productivity and save time on the lines, also because artificial intelligence is increasingly affordable and easy to use.

Certainly, Italy at system level is still far behind. 'We are a long way off,' Gatti goes on to explain, 'from the big investments of the American big tech companies that have hired the great researchers from scientific universities, or from Canada, which has created an integrated development centre with small companies. Our government has put NRP funds on the table 'but with every change of executive, development policies change, making them uncertain and weak'.

This ballast is all the heavier when one considers that, according to American forecasts, artificial intelligence has technological growth margins that extend the research spectrum for at least the next fifty years.

Someone then does not rule out the hypothesis that this artificial intelligence could be a bubble ready to burst. After all, while huge investments are calculated at 600 billion, what we can call revenues are around 200 billion. A delta that can only be filled if entrepreneurs increase their demand.

And if the small ones struggle, it is the big companies that have made the first steps forward.

This is the case, for example, for Leonardo, which decided to build its own supercomputer for data management instead of outsourcing. Michele Ippolito recounted how in this way it was possible to completely digitally redesign the fuselage of the Atr aircraft, "a successful model but dating back to the 1980s, which in this way was able to be reborn thanks to the latest technologies. Especially in a very sensitive sector like aeronautics, the possibility of creating a digital twin of the product made it possible to raise the safety bar decisively'.

Another case study is Danieli: the company that produces machines and lines for steel processing inaugurated not more than a year ago a production platform entirely managed by machine learning in all its functions, without the use of manpower. 'It is a mini scrap steel mill,' explains Anna Mareschi Danieli, 'powered by an electric furnace that is capable of autonomously separating materials, crushing them, and managing the heat and cooling chain. And all this is already a reality in Italy, near Udine'.