Artificial Intelligence, Made in Italy needs a quantum leap on data

As always, our country - accustomed to Guelphs and Ghibellines - is divided between catastrophists and techno-optimists: the former see Ia as a tool for the harmful replacement of human labour as well as a dangerous violation of common ethical principles; the latter are advocates of a massive and ungoverned use of Ia to support activities and processes in the corporate value chain. Neither perspective is really convincing; it may therefore be useful to reflect on the opportunities and risks associated with Ia with specific reference to our country's enterprises, which, as is well known, are characterised by a marked specialisation in manufacturing.

One misunderstanding must be cleared up straight away: Ia, as a tool that simplifies and supports human actions, is a determinant of structural and irreversible change in the socio-technical system. The deployment of Ia systems must, in other words, be considered in the same way as the introduction of electricity: representing, on the one hand, its pervasive impact and, on the other hand, its potential transformative effect on the dynamics of the functioning of people, businesses and objects.

Made in Italy, however, presents some very peculiar characteristics, which must be taken into account when reflecting on the potential of Ia. In particular, it is important to refer to the material legacy of our top management, which is sublimated in the affirmation of a perspective of product and production excellence as a key determinant of competitive advantage. It is no coincidence that Italy is perceived as the home of the 'beautiful and well-made'. It is a positioning that our industrial system has managed to achieve thanks to the convergence of an ultra-millennial cultural heritage, which has given rise to a creative and systemic approach that is almost unique in the world, and an ability to steer technology into manufactured products that fully meet market needs. Without risk of contradiction, we can affirm that a large part of the competitive advantage of Italy's industrial system is the result of the stratification over time of arts, crafts, and sensibilities that are not always fully codified and, as such, cannot be adequately traced back to available and fungible (bases).

It is in this perspective that we grasp the risk that Italian manufacturing runs with the large-scale diffusion of Ia. A risk that is the result, on the one hand, of the relative impermeability of our socio-technical system to digital technologies and, on the other, of our manufacturing DNA. Let me try to explain. With reference to the first point, Eurostat tells us that we are fourth last in the European Union in terms of the level of digital skills: to affirm, a relative reticence of our companies to the whole body of intangible technologies based on a perspective of connecting objects, resources and skills. The direct consequence of this state of affairs is that our companies still lack, among other things, an adequate culture of data and their protection: situations are common in which data are not collected and/or stored on a systematic basis, their quality is inadequate, or separate silos (of data) coexist in the same company that, if properly integrated, could represent a very significant source of value. On the other hand, our manufacturing legacy leads us to attribute a central and discriminating role to the production activity, i.e. the cycle of transformation of (physical) raw materials. Artificial intelligence is in itself a (new) production system that triggers a different transformation cycle, which receives data as input and transforms it into information (useful for decision-making processes) and/or action (to be transferred to robots and the like). Read in this perspective, the risk is therefore that the production-material legacy inherent in our made in Italy will become an impermeable factor towards this new transformative determinant of the dynamics of competition between companies. If we have little data or it is of poor quality, how can we think of generating Ia applications capable of sustaining the competitive advantage that we have built up over the decades thanks to productive excellence? In even more explicit terms, we could witness in the coming years the loss of this distinctive trait of ours because we do not have adequate raw material (data) to feed the production system (the AI applications) capable of making a real difference.

For these reasons, it is absolutely important that our companies take a conscious and pro-active attitude towards Ia. Firstly, by investing in effective data collection and storage systems; secondly, by devoting attention, people and financial resources to the development of Ia applications capable of qualifying a data transformation and management cycle consistent with emerging competitive dynamics; and finally, through adequate protection of the new operational cycle based on AI, which to all intents and purposes takes on a role analogous to the (more traditional) production cycle. In short, AI must increasingly become a primary source of differentiation in the value proposition of our companies and, in this framework, cybersecurity management is not only a technical action but rather a key driver for the defence of competitive advantage.