Italian Strategy for Artificial Intelligence 2024-2026

Regulatory bodies introduction

Italy's AI strategy 2024-2026 updates its 2021 program to reflect changes in technology, regulation, and generative systems, pursuing three macro-objectives: supporting AI application creation and adoption in business and public administration using country-specific data, promoting scientific research and AI applications for social wellbeing in welfare, environment, education, and health, and creating favorable conditions through training and improved public administration services. The strategy balances near-term adoption of existing AI solutions with longer-term investment in basic and applied research, addressing interconnected actions across research, public administration, business, and training sectors while considering geopolitical importance, ethical implications, privacy, security, gender balance, environmental sustainability, and rapid technology evolution.

Context analysis

Italy has a strong academic foundation in AI with 160 curricula at 53 universities and participates in major European research initiatives through institutions like the FAIR Foundation, but faces significant gaps in ICT graduates, digital skills adoption among citizens, and alignment with labor market demands. Private sector AI investment in Italy remains limited, with only 15% of small and medium enterprises initiating AI pilots in 2022 and approximately 600 patents and 350 startups since 2017, though generative AI could potentially increase national GDP by up to 18.2% annually. Public sector support includes funding for eight national competence centers extended until 2025, 40 European Digital Innovation Hubs providing subsidized digitalization services to small and medium enterprises and public administration, and various AI applications such as the €57.57 million AGENAS healthcare platform, PRODIGIT for tax justice, INPS generative AI chatbot services, and Revenue Agency anti-evasion software.

Principles of the strategy

Italy's AI strategy envisions artificial intelligence as a tool to accelerate national development across business productivity, public administration efficiency, healthcare, education, and quality of life while ensuring national security, requiring Italy to consolidate its scientific positions globally, strengthen its industrial tradition, and act as a conscious participant in technological transformation. The strategy prioritizes building domestically-developed AI technologies supported by investments in foundational and applied research, computing infrastructure, and public-private partnerships to enhance competitiveness while supporting social welfare, environmental protection, education, health for vulnerable populations, and managing demographic decline. Priority sectors include Made in Italy industries such as manufacturing, agri-food, furniture, clothing, tourism, chemicals, pharmaceuticals, and aerospace; digital and financial industries; health; education; environmental and cultural protection; privacy and security; and public administration.

Research

Italy must increase funding for foundational and applied AI research tailored to national needs and European principles of responsible AI. The strategy prioritizes public research institutions and state-funded research that can operate freely in experimentation and openly share datasets and trained models.

Public administration

Public administrations must adopt AI technologies to streamline processes and improve service quality to citizens. The strategy calls for guided implementation considering current skill gaps, targeting relevant domains for automation, decision support, and resource management through secure national platforms.

Enterprises

Italy must support business adoption of AI solutions to enhance existing processes and enable new growth opportunities. The strategy addresses skill and infrastructure barriers affecting all business sizes and emphasizes building connections between academic research and industry, with particular attention to sectors central to the Italian economy.

Education

Italy should expand high-quality teaching of artificial intelligence across universities with interdisciplinary approaches, establish specialized doctoral programs to develop researchers, and create learning initiatives for schools and citizens to build foundational knowledge. The strategy recognizes different timeframes: research requires medium-long term development, public administration and enterprises need short-term solutions, while training encompasses both sustained broad knowledge-building and urgent reskilling programs.

• the risk of “non-action”

Inaction poses a strategic threat where passive adoption of external technologies without domestic capability development wastes resources and causes competitive decline. Italy must pursue ambitious investments in AI applications tailored to national needs, values, and rights within public and production contexts, developed for Italian specificities and institutional systems, with emphasis on internal capability development over importing solutions. Geopolitical competition requires Italy to participate as a protagonist alongside the European Union rather than pursuing unfocused initiatives.

• the risk of homogenization

Widespread adoption of foreign-developed AI systems risks embedding external values and perspectives into Italian society, potentially conflicting with national constitutional values and democratic principles. The strategy must therefore promote development of generative AI models aligned with Italian and European legal and ethical standards, particularly for public administration services.

• the risk of national over-regulation

Within the complex existing European regulatory framework covering data protection, digital markets, and services, the incoming AI Act will establish harmonized rules. Italy should implement the AI Act through guidelines and flexible approaches for businesses and citizens, rather than creating additional national regulatory layers.

• the risk for the world of work

Artificial intelligence adoption will transform employment by replacing existing skills with new ones. Strategic actions must guide this transition through attention to workforce development, including upskilling and reskilling programs, to preserve and enhance work quality during the technology transition.

• the risk of the digital divide

Italy's existing gaps in digitalization and digital skills across regions risk widening if an ambitious AI strategy is not carefully managed. Training and corporate development initiatives must be systemwide and accessible nationwide rather than localized, with transparent communication about their importance and accessibility.

• the risk of ineffectiveness

The strategy must mitigate the risk that implementation fails to achieve objectives, especially given rapid artificial intelligence developments and the need to manage multiple concurrent action lines. Effective coordination mechanisms and monitoring systems with specific indicators and targets for each action are essential to adapt as conditions change.

Strategy design

Implementing the strategy's ambitions across the four priority domains (Research, Public Administration, Enterprises, Training) demands substantial organizational coordination and integrated governance. The strategy establishes an enabling framework of infrastructural and operational actions that establish common guidelines and facilitate knowledge sharing between academic institutions, industry, and government bodies, ensuring coherent deployment of initiatives across all four areas.

Training infrastructures

Italy must invest in physical computing resources for AI development and establish shared datasets and models registries to accelerate technology development and reflect national needs. The strategy requires establishing a dedicated entity with sustained funding, responsible for managing infrastructure, coordinating ecosystem actors, and monitoring implementation progress.

Public administration

This section outlines six strategic actions for AI adoption in public administration: issuing guidelines for AI adoption, establishing procurement guidance, providing development guidelines for AI applications, simplifying services for citizens and businesses, improving administrative efficiency, and integrating AI education into public administration training.

Enterprises

This section describes five strategic actions to support AI in the business sector: providing support mechanisms for small and medium enterprises, backing development and adoption of AI solutions, establishing laboratories for industrial AI applications, fostering AI startup creation, and offering specialized services for ICT companies.

Training

This section lists seven strategic actions for building AI competencies: introducing AI in schools, enabling mobility for training purposes, expanding AI teaching in university programs, strengthening national doctoral education in AI, launching upskilling and reskilling initiatives for workers, providing education on AI tool usage, and developing specialized courses in technical institutes.

A.1 | national knowledge assets: datasets and modelsi a.2 | network infrastructures for artificial intelligence

The strategy will establish a national registry of datasets and models designed for transparency, fairness, and ethical reliability, requiring all publicly funded AI projects to report their datasets and models with defined access levels, while supporting enterprise initiatives on data sharing and public administration efforts to standardize procedures and catalog public sector databases with quality and confidentiality annotations. To address infrastructure risks from widespread AI adoption, the strategy promotes intelligent networks adapting dynamically to demand through public-private partnerships, interoperable solutions, network standardization, and organizational awareness of responsible AI practices including algorithm optimization, data compression, and edge computing.

A.3 | foundation for artificial intelligence

A dedicated foundation under the Prime Minister's office will coordinate strategy implementation, manage the registry, oversee funding, and sustain operations through potential revenues. The foundation will identify development priorities, promote awareness of AI opportunities, certify AI skills, support regulatory compliance, and monitor overall strategy progress.

Strategy for research

This section outlines objectives to invest in foundational AI research while consolidating the Italian research ecosystem, and to enhance applied research through public-private partnerships. Italy must support competitiveness via international collaboration, mobility, talent retention and attraction, funding for interdisciplinary and blue-sky projects, and research emphasizing ecological, social, ethical and legal sustainability aligned with Italian and European values, alongside multidisciplinary research directly serving enterprises and productive excellence.

R.1 | consolidation of the italian research ecosystem

Strengthens a distributed network of AI research institutions, universities, and enterprises through the FAIR Foundation and extended partnerships, with focus on inclusivity and public-private collaboration. The strategy emphasizes involvement of companies that develop or benefit from AI systems, and targeted research activities serving public administration through technology validation and ethical-legal assessment.

R.2 | retain and attract talent

Requires an extraordinary hiring program across universities, research bodies, and companies to absorb talented researchers trained under PNRR funding and prevent their exit from the ecosystem. The approach includes tenure-track positions in AI fields, targeted incentives to recruit foreign researchers, and encouragement for enterprises to hire AI-trained personnel.

R.3 | design of italian large molecule models (lmm)

Establishes competitive funding for developing large language and multimodal models tailored to Italian applications in areas like public administration and health, with requirements for transparent training data, safeguards against false content generation, author rights protection, watermarking, and environmental sustainability. An ethics committee must supervise each initiative to identify risks and ensure safety and reliability.

R.4 | synergies: interdisciplinary projects for social wellbeing

Supports medium-small collaborative research projects combining multidisciplinary expertise to advance foundational science and applied AI research, evaluated on excellence and originality. Application domains include environmental protection, copyright and disinformation challenges, ethics, socio-economic impacts, AI for science, security, and cultural heritage digitization.

R.5 | AI grand challenge: foundational and blue-sky research for next-generation AI

Launches a bottom-up research program modeled on North American challenge structures to fund foundational and breakthrough research that can generate transformative technological innovation. The two-phase selection process allows researchers to propose challenge themes, with winning proposals selected from small consortia.

R.6 | enhancement of international collaborations

Two-year competitive programs will fund research initiatives designed to build cooperative relationships with European and non-European partners, aiming to develop new AI applications, harmonize adoption methods, establish protective legal frameworks, and create cybersecurity platforms. These programs will support researcher exchanges and produce international reference materials to strengthen institutional partnerships among participating nations.

Strategy for public administration

The strategy aims to support administrative processes and align services with citizen needs through structured, coherent AI adoption that establishes quality, privacy and security in data management; AI-based tools for interoperability and trust in decision-making; personnel training on AI compliance; systematic monitoring and performance measurement; and strategic decision support for evaluating AI tools, while enabling individual administrations autonomous innovation choices. Public administrations must counter discrimination in AI systems, protect intellectual property rights, and develop tools that empower citizens through accessible public services, and the strategy funds national-scale pilot initiatives and interoperable solutions while maintaining privacy and transparency. The strategy also prioritizes software neutrality to avoid dependence on proprietary platforms and to manage public resources better.

Pa.1 | guidelines for promoting the adoption of AI in public administration

Requires creating guidelines to guide public sector organizations in deploying AI, covering technological capabilities, risks, ethical frameworks, and standardized functional requirements. These guidelines will use case studies to demonstrate best practices and ensure consistent service delivery across regions and municipalities.

Pa.2 | guidelines for procurement in public administration

Requires developing guidelines focused on AI procurement to help public sector bodies select appropriate solutions through tender procedures. The guidelines must address functional requirements, security standards, and compliance with AI regulations applicable to public administration.

Pa.3 | guidelines for the development of AI applications in public administration

Requires establishing guidelines for public sector development of proprietary AI solutions using Italian platforms, with mandatory compliance to data protection and risk management rules. Guidelines must support training programs that create certified, scalable, and nationally transferable development processes.

Pa.4 | simplification for citizens and businesses

Requires designing and implementing AI-based tools and services to improve how the public accesses administrative services, potentially using foundation models, form automation, and robotic process automation. Implementation may proceed in pilot phases with careful reliability assessment before nationwide rollout.

Pa.5 | efficiency enhancement of public administration

Requires identifying and implementing AI systems to streamline internal public sector operations, such as document automation, compliance verification, and contract management. Systems must operate across the entire country and involve mapping existing implementations to assess impact and risks.

Pa.6 | AI in schools for public administration

Requires establishing an AI department within the National School of Administration offering postgraduate training for public employees at appropriate levels. Similar units should be created in all public administration training institutions to build workforce skills in AI.

Strategy for enterprises

Italy's enterprise strategy aims to strengthen the role of Italian information technology companies as developers of AI solutions and applications, and to ensure small and medium-sized enterprises adopt AI innovations to enhance their competitive position. The strategy includes funding the AI ecosystem for SMEs, fostering collaboration between enterprises, universities, and research institutions, creating international expertise networks in specific technical areas, and establishing innovative AI-focused startups, while addressing the need for businesses to transform their operations through data and AI technologies. The strategy intends to support Italian enterprises by financing an AI-focused ecosystem that strengthens their distinctive excellence through technological solutions, increase access to funding for new AI project initiatives, and enhance the AI startup ecosystem by attracting both public and private capital.

I.1 | AI facilitators for SMEs

Requires establishing territorial hubs managed by the AI Foundation that connect technology providers with businesses in specific sectors like food, textiles, and aerospace. These facilitators assess company readiness, enable solution adoption, collect training data, certify competencies, and support startup development.

I.2 | support for the development and adoption of AI solutions

Requires creating funding mechanisms supporting both technology companies developing new AI solutions and non-technology businesses adopting them. Financing options include direct loans, venture capital, private equity, and vouchers for innovation support services through designated facilitators.

I.3 | laboratories for the development of AI applications in industrial contexts

Promotes creation of thematic laboratories and centers for applied AI that bring together companies, universities, and public research entities with equal contributions. These facilities will conduct exploratory research on market-focused technologies, enable supply chain solutions, bridge institutional and international research gaps, support training for AI professionals and technicians, and assist with certification and patenting.

I.4 | development of start-ups in the AI sector

Addresses Italy's lower number of AI start-ups compared to other European countries by promoting consolidation of incubation initiatives, supporting university spin-offs, and establishing AI-Enterprise labs within universities. Aims to foster interconnection between start-ups and established enterprises through networking, public-private partnerships, sandbox projects, and international promotion of AI solutions.

I.5 | services for ICT companies on AI

Italian companies face regulatory compliance costs that hinder market participation in the rapidly evolving AI sector. The government will support businesses, particularly smaller firms and new ventures, through funding mechanisms or advisory services to navigate certification requirements, enter regulatory experimentation zones, and achieve compliance with the AI Act and sector-specific regulations such as EASA standards for aerospace.

Training strategy

Italy's training strategy aims to develop high-level professional AI skills through higher technical institutes, universities, and research doctorates while addressing gender gaps and the multidisciplinary nature of AI, and to establish reskilling and upskilling programs so workers remain competent as technology changes. The strategy also promotes broad AI literacy across schools, workers, and society, particularly vulnerable populations, through school awareness programs, public information campaigns, and fostering understanding of AI's ethical and social implications to prevent digital divides and maintain social and economic cohesion.

F.1 | pathways for introducing AI in schools, students and teachers

Requires gradual introduction of AI education in primary and secondary schools through teacher training first, followed by student pathways emphasizing AI literacy, practical tool use, and ethical awareness. Calls for nationally coordinated training delivered in multiple formats with shared educational materials and oversight by technical experts and education specialists.

F.2 | mobility as a tool for AI training

Establishes four mechanisms for exchanging skills and knowledge: funded internships for university students in companies, secondments between universities and enterprises, research mobility grants for knowledge transfer among institutions, and dual high education apprenticeships allowing young people to combine training with employment.

F.3 | widespread AI education in university degree programs

Universities should integrate foundational AI teaching into all study programs, including non-technical disciplines, tailored to each field's goals. Institutions will be encouraged through performance-based funding mechanisms based on the number of programs offering such integrated pathways.

F.4 | enhancement of the national phd program in AI

The national PhD-AI program must receive sustained funding of at least 40 scholarships annually per doctorate across five programs to maintain this strategic doctoral training beyond its initial funding cycle. Educational activities coordinated by the program should include international guest teaching, doctoral schools, and events to disseminate research results.

F.5 | upskilling and reskilling programs for businesses and public administration

Upskilling and reskilling programs must adapt training to help workers transition into jobs created or transformed by AI technologies, with tailored content for different roles that emphasizes AI augments rather than eliminates work, particularly targeting sectors where AI has matured and workers facing adoption challenges. Training must address management staff to reorganize operations around AI capabilities and technicians in both public administration and businesses to update technical competencies, modeled on successful initiatives like Elements of AI to ensure wide accessibility. Industry-specific academies structured as consortia involving training organizations, industry associations, and medium-to-large companies must design and deliver courses for their own workers and those of small and medium enterprises in their supply chains while attracting young talent.

F.6 | education on the use of AI tools

Multiple national actors should conduct coordinated awareness campaigns through diverse channels, including media programs, social media, events, and public debates, to help citizens understand how to use AI tools responsibly. A technical-scientific committee will centrally coordinate these initiatives to define scope, channels, and messaging.

F.7 | its courses focused on AI

Higher Technical Institutes should develop and integrate AI courses into their two- to three-year training programs for secondary school graduates, in collaboration with universities and specialized ICT companies developing AI solutions.

Monitoring the strategy

Each strategic action has target values measured by key performance indicators, with the Foundation for Artificial Intelligence conducting overall monitoring and producing annual reports evaluating each strategic area. The Foundation uses a panel of experts applying the Delphi method to analyse KPI and flagship project data: for education, monitoring tracks interdisciplinary AI degree courses, doctoral student enrollment and public AI awareness; for research, it measures scientific publications, strategic initiatives and startup creation in AI. Public administration growth is evaluated through two flagship projects measuring improvement in citizen service access, digitalization and streamlining of administrative procedures, while business sector monitoring measures AI adoption in SMEs and initiatives targeting SMEs involved in AI technology provision.

Regulatory bodies

Italy will establish an Agency as its supervisory authority under the European AI Act, which uses a risk-based approach with varying compliance obligations depending on risk levels; the Agency must support those developing or implementing AI in understanding European and national legislation, provide guidance to the production system, perform supervisory and notification functions to ensure AI Act compliance by manufacturers and distributors, and verify certification activities for high-risk AI systems. The Agency must coordinate with entities responsible for data protection, consumer protection, digital services, banking and financial regulation, and the National Cybersecurity Agency while remaining independent in its distinct functions.