­­Meeting the AI Act: Building the ecosystem of European AI literacy through cross-national policy engagement and cooperation 

The European Union’s Artificial Intelligence Act (AI Act) of 2024 marks a shift from principle to practice in setting new standards for AI governance. The transition, however, from regulation to operational reality remains challenging, particularly in the case of AI literacy. Indeed, the AI Act sets expectations for a company’s staff to have a “sufficient level of AI literacy” in article 4. The term “sufficient level” is quite broad and may benefit from clearer guidance, as it can be challenging for many organisations – particularly those with limited resources, such as SMEs and start-ups – to interpret and apply effectively. During Artificial Intelligence Skills Alliance’s (ARISA) Final Conference on 26 March 2026, the panel discussion “From AI literacy to AI deployment: operationalising the AI Act across Europe” delved deeper into how policymakers may make the abstract concrete.  

Aligned with this objective, the ARISA project has supported cross-sector collaboration since its launch in 2022 to advance AI skills and ensure that European recommendations are adapted to national contexts for long-term development. AI literacy is a core challenge for Europe, with the skills gap still wide. Yet companies, particularly small businesses that would benefit from efficiencies promised by AI systems, continue to struggle with implementing the technologies and accompanying regulations. The development of an AI skills ecosystem that brings together policy and decision makers, industry and education is paramount to clarify a field that remains abstract for many. The ARISA Alliance constitutes a major step in this direction and works to develop the tools to enable long-lasting policy change for an inclusive, future-oriented skill ecosystem. 

Though the AI Act was designed to manage risk, it quickly became clear that without the necessary skills, implementation and compliance with the regulation would be difficult. Article 4 covers that aspect, requiring organisations to ensure basic AI literacy from staff and even to end-users. AI literacy does not just signify awareness, but encompasses the practical capabilities needed to deploy, manage and assess AI systems. However, the ambiguity persists: What does “sufficient” literacy as stipulated in the legislation look like? How should it be demonstrated and who is accountable when AI systems span multiple stakeholders? Experts of the panel stated that enforcement will likely focus on “obligation of means” rather than results, at least in the short term. Yet, the absence of concrete benchmarks risks leaving organisations in a state of uncertainty. 

Indeed, for many companies the path to AI literacy knows several barriers. First, there is a struggle to articulate AI and strategic goals, which leads to hesitation in investment and adoption. Second, we can note a critical shortage of hybrid profiles. On top of technical expertise, there should be domain experts to bring a mix of skills. Third, fears around AI and its potential to change the job market as we know it makes employees distrustful of this technology and resistant to its implementation. It then appears imperative to work on transparent AI. Lastly, companies are facing difficulties embedding AI into existing systems and governance structures. These issues must be addressed and to do so there is certainly a need for technical training, but also change to strategies in order to build trust and show AI’s value. 

So, Europe’s AI skills challenge is not merely a shortage problem, but a composition problem. While technical skills are abundant, there is a need of integrated, product-oriented, and interdisciplinary capabilities. Innovation in AI is not about invention alone, rather it is about transforming technical knowledge into tangible impact. Panelists discussed how educational institutions and training programs must evolve to bridge the gap between academic knowledge and real-world application. For this, programs should be designed to be able to quickly adapt to market demands and emerging technologies. Universities should focus on foundational and transferable skills, while non-university providers deliver on applied, industry-relevant training. Nevertheless, it is not accepting defeat to admit that to some extent we will always be behind in skills given AI’s rapidly ever-evolving nature. 

Fragmentation between EU-level guidance, national initiatives, and industry efforts constitutes a major risk. As such, policy coordination is imperative. EU guidance should be clearer, with explicit benchmarks and coordination mechanisms to ensure consistency. Localised initiatives are also very valuable, as demonstrated during the discussion by Slovenia’s AI competence centre, which acts as a coordinator and provides hands-on support to businesses. Europe should encourage regional coordination for tailored approaches leveraging regional expertise. Decision makers should, for their part, clarify expectations by providing concrete guidance on AI literacy requirements and enforcement mechanisms; invest in hybrid skills; enhance collaboration between all levels (European, national, individual stakeholders) to avoid fragmentation; tailor skills initiatives to economic contexts; and promote trust and transparency by addressing workforce concerns. 

Though education is mainly a competence reserved for the EU Member States, with the European Union playing an important supporting role, members of the skills ecosystem – companies, education providers, policy and decision makers – can find their own path to achieving policy objectives. Several routes towards implementing best practices can be followed: 

• Building AI literacy strategically: Begin by assessing current knowledge, defining clear learning objectives, and providing targeted training programs to support staff understand AI fundamentals, applications to their work environments, risks, and regulatory requirements. This ensures that organisations not only comply with the regulation but also build foundational skills critical for safe and effective AI deployment.  

• Domain-focused skills: Prioritise skills that drive tangible innovation and sector-specific performance. Rather than focusing solely on general IT competencies, develop expertise that integrates digital capabilities with industrial, scientific, entrepreneurial, and sector-specific knowledge.