Artificial Intelligence as a Driver of Economic Growth and a New Source of Productive, Employment, and Territorial Polarization in Italy’s Digital Transition
Abstract: The Governor of the Bank of Italy’s Final Considerations and the 2025 Annual Report place artificial intelligence at the centre of contemporary economic transformation, not as a mere sectoral innovation, but as a general-purpose technology capable of reshaping production, labour, investment, competitiveness, and firm organization. AI has already entered macroeconomic dynamics, supporting investment, trade, and financial valuations, but its diffusion shows strong geographical and productive asymmetries: the United States maintains a dominant position in frontier models and computing capacity, China is rapidly closing the gap, while Europe and Italy risk falling behind if they fail to turn technological adoption into widespread productivity gains. For the Italian economy, the decisive issue is not only the use of AI by large corporations, but its penetration into small and medium-sized enterprises, manufacturing, professional services, healthcare, logistics, and administrative processes. Data cited by the Bank of Italy show still-limited but growing adoption, with expected effects more pronounced on productivity than on overall employment, while the risk remains that benefits will concentrate among more highly skilled workers and more structured firms. Italy’s challenge is therefore to transform artificial intelligence from a selective source of competitive advantage into a national lever for growth, training, productive inclusion, and labour renewal.
Keywords: #ArtificialIntelligence #BankOfItaly #FabioPanetta #AnnualReport2025 #ItalianEconomy #Productivity #Labour #Firms #SMEs #Innovation #DigitalTransition #Competitiveness #Employment #Training #Inequalities #EthicaSocietas #SocialSciences #EconomicPolicy #FrancescoMancini #EthicaSocietas #EthicaSocietasJournal #ScientificJournal #SocialSciences #ethicasocietasupli
Artificial intelligence as a general-purpose technology of economic transformation
The Governor of the Bank of Italy’s Final Considerations, presented today at Palazzo Koch on the occasion of the 2025 Annual Report, assign artificial intelligence a role that goes far beyond that of a simple technological innovation, as it is placed within a broader transformation of production systems, labour relations, investment patterns, and the growth trajectories of advanced economies. The Governor highlights that AI is not a sector-specific application aimed at improving individual business functions, but rather a general-purpose technology, comparable in scope to the major innovations that have historically reshaped how societies produce, organise labour, allocate capital, and define their position in international competition.
The central issue for Italy is not whether artificial intelligence will enter economic processes—because this is already happening—but whether the national productive system will be able to transform it into widespread productivity growth, or whether it will instead deepen existing divides between large and small firms, between highly skilled workers and those exposed to task displacement, and between digitally advanced and structurally weaker territories. The issue is therefore not only technological but fully economic, social, and institutional, as it concerns the country’s ability to govern a transition that can generate growth only if accompanied by coherent policies of training, investment, technological transfer, and productive reorganisation.
Governor Fabio Panetta stressed that artificial intelligence has already entered macroeconomic dynamics, supporting investment, trade, and financial valuations, but also emphasised that its impact is broader, as it progressively reshapes the way production, work, and decision-making are carried out. AI, therefore, cannot be considered a marginal variable of productive modernisation, but must instead be interpreted as one of the key infrastructures of future growth, with positive effects achievable only if technological adoption is accompanied by organisational and cultural transformation within firms and public administrations.
The issue of Italian productivity
In the past, the Governor has highlighted how low productivity is a major problem of the Italian economy, limiting real wage growth, reducing firms’ competitiveness, restricting both public and private investment capacity, and making it more difficult to sustain welfare, innovation, and the country’s financial stability in the long run. Artificial intelligence may represent a potential lever of disruption in productivity, as it can reduce costs, speed up decision-making, improve design processes, optimise maintenance, support research activities, make service organisation more efficient, and enable better resource allocation.
However, productivity does not arise simply from the availability of technology or the purchase of isolated digital tools, but from the ability to integrate innovation, human capital, data quality, and process reorganisation. An AI system embedded in an inefficient organisation risks automating pre-existing inefficiencies, just as an algorithm fed with incomplete or poorly structured data may produce fragile, distorted, or difficult-to-verify decisions. Therefore, AI’s contribution to Italian productivity will depend not only on its diffusion, but on the depth with which firms are able to rethink tasks, procedures, information flows, internal responsibilities, and the relationship between human labour and cognitive automation.
Artificial intelligence can help close Italy’s productivity gap only if it is adopted as part of a productive strategy rather than as a marginal efficiency tool. Firms that use AI merely to reduce time or compress costs will achieve limited benefits, whereas those that integrate it into design, logistics, predictive maintenance, customer management, data analysis, research, and internal training can turn innovation into a stable competitive advantage.
The international concentration of technological power
The Final Considerations highlight a particularly important fact: the development of frontier models and computational capacity is currently highly concentrated in a few international actors, with the United States in a dominant position, China rapidly closing the gap, and Europe still significantly behind. This has not only industrial significance but also strategic implications, as control over models, data, computational infrastructure, talent, and platforms is shaping a new geography of global economic power.
For Italy and Europe, the issue is not merely closing a technological gap, but recognising that artificial intelligence affects productive sovereignty. An economic system that relies predominantly on externally developed technologies can certainly gain application benefits, but risks becoming dependent on architectures, standards, costs, governance models, and industrial decisions it does not fully control. Technological dependence thus becomes economic dependence, as it constrains firms’ capacity to innovate, territories’ ability to compete, institutions’ regulatory power, and citizens’ effective control over data usage.
However, the Bank of Italy notes that in major technological revolutions, the greatest benefits have not necessarily gone to those who originated the innovation, but often to those able to adopt, adapt, and apply it more effectively. This is particularly relevant for Italy, as it suggests a realistic strategy: rather than focusing solely on building global foundation models—currently dominated by a few international players—Italy and Europe should develop a national capacity for high-quality AI application in sectors where the country has distinctive strengths, from advanced manufacturing to healthcare, professional services, logistics, agrifood, culture, public administration, and tourism.
The digital transition of Italian firms
The Annual Report shows that the basic digitalisation of Italian firms has now reached significant levels, with a very high share of companies equipped with a minimum level of digital technologies. This is an important but insufficient precondition for AI adoption. Basic digitalisation refers to the presence of tools, connectivity, and IT procedures, whereas AI requires an additional level: the ability to collect and process data, system interoperability, cybersecurity, internal skills, information flow organisation, and willingness to modify decision-making processes.
The data reported in the Governor’s analysis show that AI adoption among Italian firms is still limited compared to the European average, but growing. In 2025, only a small share of private firms in industry and non-financial services with at least ten employees were using AI tools, while surveys of firms with at least twenty employees indicate an acceleration between 2025 and 2026, with expectations of further growth by the end of the year. This dynamic shows that Italy is not excluded from the transition, but is still in a selective phase, where adoption is concentrated in more structured firms, more advanced sectors, and companies with stronger organisational and financial capacity.
The issue is therefore no longer simply to promote digitalisation, but to prevent AI adoption from remaining confined to a narrow segment of the productive system. In a country characterised by a dense fabric of small and medium-sized enterprises—often highly specialised but not always equipped with adequate internal resources—the risk is that artificial intelligence will reinforce already competitive firms while leaving behind those that would need innovation most to recover productivity.
The size gap as an economic and social divide
The size gap is one of the most critical aspects of AI’s impact on the Italian economy, as AI diffusion is significantly higher among large firms. This is not surprising, since large firms more easily have access to structured data, internal expertise, investment capacity, dedicated innovation units, specialised consultancy, and the ability to absorb the initial costs of experimentation and organisational restructuring.
In Italy, however, this difference is particularly significant because the productive system is largely based on small and medium-sized enterprises, often embedded in territorial supply chains, industrial districts, and complex production networks. If AI becomes accessible only to larger firms, it risks widening internal competitiveness gaps, leading to a dual-track economy in which part of the country accelerates while another part stagnates or falls behind.
To avoid this outcome, economic policy must treat AI adoption in SMEs not as a private matter of individual firms, but as a national industrial policy issue. Shared platforms, targeted incentives, managerial training, technology transfer centres, business consortia, accessible digital services, credit tools, and operational support mechanisms are needed. Artificial intelligence becomes a driver of national growth only if it penetrates the real productive fabric, not just the most capitalised segments of the economy.
Work and transformation of tasks
The impact on work deserves specific consideration, because artificial intelligence has a distinctive feature compared to many previous technologies: it is capable of performing tasks with a high cognitive content, affecting activities that until a few years ago were considered relatively protected from automation. This does not mean that human work is destined to disappear, but rather that the composition of tasks and the relative value of skills will change.
Governor Panetta recalled that in history all major technological innovations have made some professions obsolete while simultaneously creating new ones. Artificial intelligence may follow the same trajectory, especially if productivity growth reduces costs and prices, expands demand, supports economic activity, and fosters employment. However, the transition will not be without costs, as not all workers will be able to easily move from declining activities into emerging ones.
The labour market is likely to evolve in two directions: on the one hand, AI will support people in repetitive, analytical, or document-based tasks, allowing them to focus on higher value-added functions; on the other hand, it will make even more relevant those skills that remain distinctly human, such as interpreting results, exercising judgment, taking responsibility, ensuring process reliability, understanding complex contexts, negotiating meanings, and evaluating the ethical and social consequences of decisions.
The risk is that the benefits of this transformation will be concentrated among more highly skilled workers, while those with weaker skills or more exposed tasks may experience a loss of professional centrality. AI therefore creates not only an employment issue, but above all an issue of job quality, distribution of opportunities, and social justice in the transition.
Training and human capital as conditions for growth
For artificial intelligence to become a driver of broad-based growth rather than a factor of polarisation, training must be considered an essential economic infrastructure. It is not enough to introduce technological tools into firms if workers, managers, professionals, and administrators are not equipped to understand them, use them critically, integrate them into processes, and manage their limitations.
The training required by AI does not coincide with generic digital literacy, but involves deeper skills: the ability to work with data, understand the logic of automated systems, assess the reliability of outputs, recognise errors and distortions, integrate algorithmic suggestions with professional judgment, reorganise processes, and take responsibility in hybrid human–machine environments. This is where a significant part of Italy’s competitiveness will be determined.
Public policy should therefore focus on continuous reskilling programmes, strengthening higher technical education, collaboration between firms and universities, support for training in SMEs, and accompanying tools for more exposed workers. If training is left solely to individual initiative or to the capacity of stronger firms, AI will increase inequalities; if instead it becomes a national policy, it can turn into a tool of productive inclusion.
Prices, margins, and distribution of value
Another aspect concerns the distribution of economic benefits generated by artificial intelligence. Productivity gains can translate into lower costs, lower prices, higher margins, better wages, or new investments, but the actual direction depends on market structure, the degree of competition, workers’ bargaining power, and institutions’ ability to steer growth towards socially sustainable objectives.
Firms adopting AI may achieve efficiency gains and, in some cases, slower price increases; in the medium term this could help ease certain inflationary pressures, but it does not automatically ensure an equitable distribution of the value created. If benefits are retained mainly by stronger actors, productivity can increase without corresponding improvements in workers’ or consumers’ conditions.
For this reason, artificial intelligence must also be integrated into a broader reflection on the relationship between productivity, wages, profits, and investment. Technological growth that only increases the profitability of a few actors without expanding social opportunities risks reinforcing inequalities; growth accompanied by training, collective bargaining, organisational innovation, and industrial policy can instead help build a new balance between economic efficiency and social development.
Public administration, data, and national competitiveness
The impact of AI on the national economy does not concern only private firms but also directly involves public administration, as the quality of public services, procedures, justice, healthcare, taxation, procurement, authorisations, and information infrastructure affects the overall productivity of the economic system. A slow, fragmented, and poorly interoperable public administration can reduce the benefits of private innovation, whereas a digital, skilled, and AI-capable administration can become a multiplier of national productivity.
Artificial intelligence can support public administration in data analysis, process simplification, anomaly detection, document management, decision support, and service quality for citizens and firms. However, the fundamental principle must remain public accountability. AI cannot replace administrative decisions nor make procedures opaque; instead, it must strengthen traceability, efficiency, and verifiability.
Italy’s competitiveness will therefore also depend on the State’s ability to become an intelligent infrastructure of the transition, not merely an external regulator. This requires high-quality public data, interoperability, cybersecurity, training of public officials, clarity of responsibilities, and governance capable of preventing administrative automation from turning into a new form of algorithmic bureaucracy.
The risk of new inequalities
Artificial intelligence can increase productivity, but it does not automatically guarantee that these gains are evenly distributed. It may strengthen larger firms, more skilled workers, already advanced regions, and more capable administrations, while leaving behind those who most need innovation to regain competitiveness and inclusion.
In Italy this risk is particularly high because existing fractures are already deep: territorial divides between North and South, size disparities between large firms and SMEs, generational gaps, digital skills lags, uneven public administration performance, and unequal access to innovation. AI does not create these inequalities on its own, but it can amplify them if adopted without compensatory policies and without a national diffusion strategy.
The role of institutions is not to slow innovation, but to prevent it from becoming the exclusive domain of the strongest actors. Governing AI means making it accessible, understandable, verifiable, and socially oriented; it means supporting smaller firms, protecting more exposed workers, training people, reducing territorial gaps, and turning technological growth into social progress.
Opportunity or risk depending on AI implementation
The Final Considerations of the Governor of the Bank of Italy and the 2025 Annual Report indicate that artificial intelligence is now a structural component of the new economy, destined to affect productivity, labour, firms, prices, investment, and the international positioning of productive systems. For Italy, it represents a concrete opportunity for relaunch, but also a test of institutional and social maturity.
The country can use AI to address one of its historical weaknesses—low productivity growth—but only if it avoids a selective modernisation that concentrates in a few firms, territories, and professional profiles. The productivity generated by AI must become widespread productivity, and this requires training, investment, industrial policy, digital infrastructure, technological transfer, and a public administration capable of accompanying change.
The real question, therefore, is not whether AI will change the Italian economy, but whether Italy will be able to govern this change. If left solely to market dynamics, artificial intelligence may increase efficiency and inequalities at the same time; if embedded in a national economic strategy, it can instead become a lever of development, inclusion, and productive renewal. It is within this alternative that the challenge identified in the Governor’s Final Considerations lies: transforming technological innovation into social progress, so that productivity does not become the privilege of a few, but the foundation of a new pact between economy, work, and society.
OTHER CONTRIBUTIONS BY THE SAME AUTHOR
DRAGHI IN AACHEN: EUROPE ALONE BEFORE THE TEST OF SOVEREIGNTY
SIGONELLA 1985: WHEN ITALY ASSERTED ITS SOVEREIGNTY
HOLOCAUST REMEMBRANCE AND CONTEMPORARY ANTISEMITISM
RAID IN VENEZUELA: THE USA CAPTURE MADURO AND HIS WIFE TO PUT THEM ON TRIAL
FIVE LATEST CONTRIBUTIONS
ARTIFICIAL INTELLIGENCE AND CLINICAL RISK IN HOSPITALS: FROM AUTOMATION TO GOVERNED ACCOUNTABILITY
DESTROYING IMAGES, GOVERNING MEMORY: JIHADIST ICONOCLASM AND COGNITIVE WARFARE
THE IRREDUCIBLE CITY: JERUSALEM IN THE AGE OF THE GLOBAL CRISIS
Ethica Societas is a free, non-profit review published by a social cooperative non-profit organization
Copyright Ethica Societas, Human&Social Science Review © 2026 by Ethica Societas UPLI onlus.
ISSN 2785-602X. Licensed under CC BY-NC 4.0