There comes a moment when a technology stops being “promising” and becomes inevitable. For Defense, that moment has arrived: Artificial Intelligence is no longer an optional choice or a niche experiment, but a strategic lever that impacts operations, organization, training, and industry relations. This is the central message of the “AI and DEFENSE – Defense Strategy on Artificial Intelligence, 2026 Edition”, which outlines a concrete and ambitious path, set on short timelines and measurable objectives similar to what the United Kingdom already did in 2022.
A Clear Vision: Integrate Quickly, but Responsibly
The strategy starts with a clear observation: contemporary competition is accelerated, multi-domain, hyper-informed. In this context, AI becomes the element that allows reducing the technological gap, strengthening Italy's credibility in alliances (NATO and EU), and above all, increasing operational effectiveness without losing sight of values.
The document emphasizes two fundamental “pillars”:
- responsible, ethical approach in compliance with International Humanitarian Law (IHL);
- human control and responsibility of the command line, because technology should amplify decision-making, not replace it.
This framework is not a detail: it is the condition to ensure that innovation is sustainable, interoperable with partners, and defensible even on political and legal grounds.
Strategic Autonomy and Technological Sovereignty: The Direction is “In-House”
The core of the “step forward” lies in the declared objective of strategic autonomy and technological sovereignty. Translated: buying solutions is not enough. Defense must maintain critical internal capabilities to:
- train/update models even with recent data and in challenging contexts;
- understand algorithms, limits, biases, and risks;
- ensure compliance with IHL, doctrine, and values;
- maintain operational systems even without external support.
This is not an “ideological” stance, but a practical response to three real problems: supplier dependencies, vulnerabilities (even intentional), and the risk of losing freedom of action in geopolitical crises or in case of corporate failures.
Enabling Assets: Talents, Data, Algorithms, Computing, and Secure Networks
Every strategy is credible if it names the material conditions that make it possible. Here, Defense does so, clearly indicating the pillars:
- qualified human resources, with attention to “talents” and their attraction/retention;
- data (integrity, quality, absence of bias, overcoming “silos” and governance);
- algorithms (development, validation, maintenance, and certification);
- proprietary computing capacity, with high requirements (also in perspective of HPC and hybrid HPC–quantum);
- secure networks and connectivity, with resilience and continuity as non-negotiable criteria.
The very operational approach is striking: computing is not an “IT luxury,” but the factor that fuels a virtuous cycle (more computing → more experiments → better models → more capacity) that determines the pace of advantage.
A “Modular” and Fast Path: 1 Year, 2 Years, 3 Years
The strategy chooses an aggressive timeline, precisely because AI is advancing rapidly. The path is divided into three horizons:
- short term (within 1 year): consolidate enabling conditions (governance, infrastructure, skills, processes, obstacle removal);
- medium term (within 2 years): implement priority operational projects;
- long term (within 3 years): widespread, secure, and sustainable integration in all areas of Defense.
This is a key point of the “decisively towards the future”: not a wish list, but a schedule that makes ambition controllable and verifiable.
Four Areas of Intervention: Where AI Truly Changes the Game
The strategy identifies four main areas covering the entire Defense machinery.
1) Operations: Decision Superiority, Engagement Speed, Multi-Domain Protection
In the modern battlespace, the mass of data and battlefield transparency make human analysis capacity alone insufficient. AI is described as a tool for decision augmentation, “adaptive” wargaming, predictive analysis, multi-source intelligence, and protection against cyber/electromagnetic threats and against drones/swarms.
The document also addresses the more sensitive implications (such as autonomous weapon systems) emphasizing the relevance of ethical profiles and governance.
2) Organizational Contexts: Productivity, Services, Processes
Here there is a concrete promise: to achieve rapid benefits. It explicitly mentions tools for individual productivity and a possible “Defense” Large Language Model trained on internal documentation, as a prerequisite for knowledge management, drafts, summaries, information extraction, and terminological consistency.
Alongside this: user assistance with natural language, integration with automations (RPA), and, in perspective, AI agents for unstructured problems — always with the rule: human judgment must not be eclipsed.
3) Training, Education, and Research: AI Culture and Digital Mindset
The idea is clear: without widespread training, adoption is fragile. It talks about structured AI literacy (cognitive, operational, critical, and ethical-normative dimensions) and reskilling/upskilling to bridge generational gaps and build a mixed human-machine workforce.
4) Industry: National Ecosystem, Open Innovation, More Agile Procurement
The document values the national industrial ecosystem (large players, but also startups and SMEs), with a shift towards partnerships, technological scouting, and more agile mechanisms for pilot projects. The goal is twofold: accelerate innovation and reduce dependencies, strengthening a competitive “Defense Tech” base.
Governance and Implementation: UIA and LIAD as the Central Engine
To avoid fragmentation, the strategy provides for multi-level governance with two mainstays:
- Office for AI (UIA): direction, supervision, strategic coherence, relations, and ethical-legal oversight;
- AI Laboratory for Defense (LIAD): center of excellence and AI Delivery Center, aimed at prototyping, experimenting, and operationalizing solutions in synergy with end users.
Decisive element: within 3 months of approval, an Implementation Plan with actions, responsibilities, and KPI monitoring is expected. This is where the strategy stops being a “declaration” and becomes management.
Conclusion: A Step Forward That Demands Continuity
“AI and Defense” does not describe a distant future: it describes a present that demands choices. The significant step forward lies in the combination of ambition and pragmatism: close objectives, explicit enabling assets, clear governance, attention to ethics and IHL, and a strong push towards autonomy and technological sovereignty.
Now the decisive word is one: continuity. Because AI is not “introduced” once and for all: it is cultivated, governed, measured, and updated. If this path is followed with discipline, consistent investments, and the ability to attract talents, then yes: it will not just be a step forward. It will be a trajectory, finally, decisively towards the future.
Comments
No comments yet. Be the first!