To create new value, processes need to be reorganised

Reducing the impact of artificial intelligence to a simple substitution of people with machines is to fail to recognise the deeper nature of the transformation currently taking place. AI almost never enters the labour market by eliminating an entire profession en masse and replacing it with software. More often, it intervenes within specific tasks, streamlining some aspects, modifying others, and altering the relative weight of the skills required and redefines the way in which a task is organised. Professions, therefore, are not simply eliminated – often within very short timeframes – but are broken down and reassembled. Some activities lose value because they become easier to automate, whilst others gain value precisely because AI can function as a complementary tool rather than a substitute.

It is by making this distinction that we can better understand why the impact of artificial intelligence on employment will depend not only on the power of the tools available, but on the way in which these tools are integrated into production processes. The same technology can be used to do the same things with fewer people, reducing time, intermediate steps and costs, or it can be used to expand services, personalise products, improve the quality of decision-making, enter markets where it was previously impossible to operate, and build businesses that would not have been sustainable without that increase in productivity. In the first case, the effect on employment will be primarily defensive and potentially reductive. In the second, it can become expansive, because technology does not merely compress what already exists but enables the generation of new demand for skills, new functions and new roles within organisations.

For Italy, this development is particularly significant. Our production system boasts widespread technical expertise, professional knowledge accumulated in the workplace, and adaptability built up within supply chains, industrial clusters and advanced services. At the same time, however, it often struggles to transform these resources into organisational models capable of integrating new technologies without reducing them to mere tools for improving efficiency. If artificial intelligence is integrated into processes that remain unchanged, the most likely outcome will be faster execution, some cost reductions and greater pressure on certain mid-level roles. If, on the other hand, it is accompanied by a reorganisation of work, it can shift people towards higher-value activities, create bridging roles between technical expertise and customer knowledge, and strengthen functions relating to data monitoring, interpretation and governance.

The question, therefore, cannot simply be whether artificial intelligence will destroy or create jobs. It will do both, to varying degrees depending on the sector, the nature of the work and the quality of the organisations. The key is to understand under what conditions the additional productivity generated by AI can be transformed into new jobs and not merely into cost reductions. This will depend on corporate strategies, the ability of education and training systems to engage with the workplace, the quality of industrial relations, and public policies in supporting transitions that have already begun. Artificial intelligence will be able to generate employment not as an automatic consequence of the technology itself, but if it is used to increase the value of work and not merely to reduce its cost.