LET’S TALK & CONNECT

Answers of Digital Italy Summit

It is a country attentive to the complex geopolitical situation, the problems of economic growth, the productivity of businesses and the efficiency of the Public Administration that presented itself at the ninth edition of the Digital Italy Summit, held in Rome Nov. 12-14.

In the opening speeches, Fei Fei Li, Co-Director of Stanford’s Human-Centered AI Institute, traced the path of AI starting from the “traditional” phase to the generative phase and outlining the prospects of transformative AI, linked to autonomous digital agents, and finally projecting the future of “spatial” AI, based on algorithms that can generate 3D shapes from single images, and that will enable us to transform the world into digital form and simulate all its richness.

Ian Bremmer ha analizzato a fondo gli impatti dell’AI sulla geopolitica. Il livello enorme di investimenti richiesti tende a favorire una concentrazione di potere mai vista finora e una progressiva simbiosi tra grandi reti oligopolistiche e funzioni di governo.

Thus, two “technopoles,” the U.S. and China, are looming, vying for technological and political hegemony over the “swing states,” the middle powers in between. Europe, with its focus on regulatory aspects not supported by an adequate critical mass of R&D investment in innovation and the AI area, risks ending up as the crock pot among the iron pots. And only within a broad and coordinated Europe-wide initiative, such as the one envisioned by the Draghi Plan, can Europe find its own role in the strategic area of AI, and Italy can capitalize on its resources and avoid ending up as one of the wagons in the slow continental train-and not among the first.

In this context, the great wave of AI is spreading, which, beyond the obvious hype aspects (we are practically at the apex of the Gartner curve), represents both a great opportunity for growth and development, as well as, if not offset by appropriate policies, a potential amplifier of economic and social inequalities.

Indeed, the long wave of AI is entering our economies and on our societies, and it is already profoundly transforming the business models of our enterprises, the services of Public Administrations, the quality of work, and our own personal lives.

“AI is much more complex and powerful than any other technology. It is not just a tool or a platform, but a transformative meta-technology, itself capable of producing tools and platforms; it is not just a system, but a generator of systems of any kind. The risk is not to overestimate it, but rather to miss the importance of the coming wave“ (Mustafa Suleyman, ”The Coming Wave”).

Turning to an analysis of our country’s situation, we can identify three main themes in the Summit debate in particular:

The gowth theme

European industry has been in decline for some time, starting with Germany, the largest partner country for our exports. Industrial production in Italy has been declining for many months, only partially offset by the expansion of services. At present, therefore, GDP growth forecasts for 2024 range between 0.6 and 0.8 percent.

In this context, the digital market is growing by 3.5 percent, and is a dual driver of growth: on the one hand due to the development of investments especially in cybersecurity ( 13.1 percent), cloud (+25.2 percent) and artificial intelligence (+35 percent); on the other hand due to productivity gains of companies using these technologies

The second growth stimulus factor comes more generally from PNRR investment, which counts for a large share of GDP growth for the next two years.

Regardless, however, of the difficulty of realizing all the planned investments by 2026 (the Digitization Mission still records the highest percentage of spending on allocated amount among all the missions in the NRP), we have registered widespread concern about what will happen next, when the General Government will have to deal with the fallout of investments in terms of current spending.

AI as a factor in increasing business productivity.

From 2004 to 2022, according to the OECD, total factor productivity in Italy declined by 2%, registering a strongly negative gap with major Western countries (+11% in Germany, +14% in the US)

There are two reasons, however, for believing that generative AI could bring productivity gains that the world has not seen in a long time, and that would be all the more essential for our economy:

• Accessibility: On the one hand, the underlying technology is accessible to a very wide audience on existing infrastructure. A user can issue instructions to a tool like ChatGPT without having to learn a special programming language.
• Versatility: On the other hand, these tools are not limited to a particular task, function, problem or area. This makes them usable in a variety of disciplines. And once a large language model is trained on a corpus of texts, for example, it can summarize a legal document as well as a medical document or an insurance document.

These two key attributes-accessibility and versatility-suggest that widespread deployment of generative AI could face fewer obstacles than previous technological advances, thus providing a real boost to the production of goods and services.

However, harnessing the full potential of AI technology and limiting any negative effects will require implementing the right combination of policies, both from a regulatory and corporate perspective.

From the “use cases” that were presented at the Summit, however, two concerns stood out:

• that the demand for new skills far exceeds the rate at which our Education system produces those skills, leading to an exacerbation of the “skills inflation” already visibly underway.
• that the development and training costs of “Foundation Models” may continue to grow exponentially, making any assumption of competition in this area unrealistic. So better to focus on sectoral and specialized applications of generative AI, and in perspective on autonomous digital agents, which can capitalize on the “know-how” of Made in Italy Companies” and better meet the cost-benefit needs of our Enterprises.

Innovation and Artificial Intelligence in Public Administration

The Summit highlighted that there have been many improvements in the area of digital services to citizens and businesses over the past period.

SPID, CIE and App IO have now reached almost all of our population. As of Dec. 4, access to IT-Wallet via the IO app will be enabled for all, allowing digital versions of the driver’s license, health card, European disability card and later the ID card and other documents to be saved within the app.

A reality such as PagoPA in the area of payments to public administrations has very few equals internationally-for example, a country of important administrative culture such as France is far from these levels.

In parallel, the process of migration to the Cloud by Public Administrations is proceeding rapidly: thanks to PNRR investments, in one year there has been an increase from 27 to 46 percent of Local PAs that have moved all or most of their applications to the Cloud).

Also in the area of AI, we found that 6 percent of Administrations are experimenting with or using such technologies: not only in the area of automation of simple functions, such as the use of chatbots to facilitate contact with citizens, but also with more sophisticated applications, such as in the case of active labor policies, through matching labor demand with skills available on the market.

However, the Summit identified some areas that still need to be properly addressed and resolved:

• The issue of interoperability of public databases is still unresolved. As long as the immense data silos of large Public Administrations remain poorly communicating, on the one hand Artificial Intelligence will only be able to unleash a fraction of its potential, and on the other hand Prof Bassanini’s dream, recently declined in the “once only” principle (the citizen must give his data to the Public Administration once and only once) will unfortunately remain as such.
• Instead, the issue of the impact of AI on the organization of work in the PA, which is experienced today as a threat to public employees, should be reversed. In 15 years, as a result of demographic desertification and the competitiveness of more attractive private sectors, the PA workforce will probably be cut in half. This gap will have to be at least 50 percent covered by the increased labor productivity enabled by AI, or else the currently available public services will be reduced. The elimination of most repetitive work will increase the quality of work and the public value produced. And we also trust that we will have public services not only similar to what we have today, but much better.