How PA must adopt artificial intelligence
Artificial intelligence today places the Italian Public Administration at a historic crossroads: it can either be the greatest opportunity for innovation ever, or prove to be yet another technology to speed up outdated procedures. The Italian PA will only be in a position to 'generate the future' if it knows how to manage the adoption of AI in a conscious and strategic manner. And to achieve this, it must not automate the existing, but rethink itself, with new processes, new skills, new organisation.
In the past, we have confused digital transformation with the simple transposition of analogue into electronic. We have digitised forms, procedures, bureaucracy and often inefficiencies without questioning processes, roles and models. The result was a PA that was more computerised, but not always more efficient. Today, the risk is to repeat the same mistake if we do not fully understand the challenge that the public sector faces with AI.
The anticipatory governance. The PA, at the end of the PNRR, is going through a crucial phase. After having successfully built in recent years new digital infrastructures, new spending capacity, new ways of making decisions, widespread skills, territorial alliances, it must make a qualitative leap. And to achieve this, it must develop 'anticipatory governance' that enables it to read the weak signals of change, imagine alternative scenarios and prepare for tomorrow.
With this in mind, artificial intelligence is an extraordinary tool because, in addition to automating repetitive tasks, it can increase the ability of administrations to interpret complexity, simulate impacts, read data, and guide decisions. For the first time in history, administrations can also use intelligent systems to govern the future. Together with computational power, they can understand the effects of decisions before they fully manifest themselves. Think about what this means in practice in different sectors, from urban planning to healthcare, environmental management, mobility or welfare: AI can help predict needs, allocate resources, identify critical issues before they become emergencies.
Cultural transformation. But all this requires a cultural transformation before a technological one. We have to stop asking just "what can AI do for public administration?" and start asking ourselves what PA we want to build with the support of AI. And this is by no means immediate. Today, all public organisations seem to be racing in the same direction: they organise prompting courses to use LLM systems, driven by the anxiety of not falling behind, often without asking where this race is leading.
