Blended AI Skills: Shaping PRODIGI’s youth AI Literacy Scale (yAILS)

By Leen d'Haenens

2026/01/12

A 16-year-old in Flanders might use an AI chatbot to finish homework, yet struggle to explain what an algorithm actually does. This gap between routine use and real understanding of artificial intelligence is what the PRODIGI project set out to explore. In a recent pilot survey of 1,200 young people in Flanders, Belgium, we tested a newly developed youth AI Literacy Scale (yAILS), revealing that teens tend to blend technical, creative, and ethical skills when interacting with AI. This early finding is now guiding PRODIGI’s next steps, from refining the survey itself to testing targeted interventions with the young people who need them most.

Groundwork: Listening to Youth and Experts in Scale Development

Before crunching any numbers, the PRODIGI team laid critical groundwork by involving those at the heart of the issue in cognitive interviews: vulnerable youth and subject experts. Specialists in AI and education reviewed dozens of draft questions, flagging any fuzzy wording or unnecessary jargon. They recommended using simpler language and cutting technical terms that might confuse many teens (for example, words like “algorithm” or “bias” can bewilder students, especially those from migrant backgrounds). The expert panel also pushed for concrete examples in questions about big ideas like responsibility or fairness, making sure abstract concepts were tied to relatable scenarios. Youth echoed the need for simplicity. In face-to-face interviews, teenagers showed curiosity about AI but often stumbled over the way questions were phrased. Many didn’t realize how much AI they already use daily, and they found technical terms like “algorithm,” “deepfake,” or “bias” perplexing. Long, academic-sounding sentences lost them. The PRODIGI team took these lessons to heart, revising the item pool for clarity and cutting any confusing or redundant questions.

A Pilot in Flanders: Five Dimensions Become Four

We launched a large-scale pilot in Flanders to see how the proposed AI literacy scale held up quantitatively. About 1,200 youngsters across the region answered a battery of questions on what they know and can do with AI: from operating AI tools to creating content and dealing with ethical dilemmas. With this trove of data, we ran an exploratory factor analysis (EFA) to test whether the skills naturally grouped into the five categories we had originally envisioned.

The results were eye-opening. Instead of the five distinct skill domains the team started with, the stats revealed four strong, clear factors. In other words, some of the conceptual buckets had fused together when confronted with real youth responses. The refined Youth AI Literacy Scale now encompasses four interrelated dimensions:

• Technical and operational skills: the hands-on ability to use AI tools and understand their basic workings.

• Content creation and communication: using AI in making new content (like images or text) and interacting through AI (e.g. chatbots, voice assistants).

• Information navigation: finding, evaluating, and managing information in an AI-rich online environment (including verifying what’s true).

• Ethical and responsible use: understanding AI’s biases/limitations and applying moral judgment when using AI (such as fairness, privacy, and accountability).

These four dimensions are both theoretically grounded and empirically coherent, meaning they make sense conceptually and were borne out by the data. The collapse from five to four happened because, as the researchers saw, young people don’t compartmentalize AI skills in practice. The EFA showed that a question intended to tap one skill often correlated with another, indicating that competencies overlapped. AI tasks require integrated, not isolated skills, as prompting, creating, evaluating, and ethical reasoning often co-occur. A teenager using AI might simultaneously employ technical know-how, creativity, critical thinking, and ethical judgment in a single interaction. The new four-factor structure embraces this reality, reflecting “hybrid” skill sets rather than siloed abilities.

Key Findings: Integrated Skills and Gaps to Fill

Beyond reshaping the scale’s structure, the Flanders pilot survey yielded important insights into what young people know (or don’t know) about AI. One striking finding was the overall knowledge gap. When confronted with 30 factual questions about AI (covering concepts like how algorithms work or how to detect AI-generated content), nearly half of the participants scored only 16 or fewer correct answers. In other words, a substantial portion of youth have less than modest AI knowledge, underscoring the need for better education in this area.

Crucially, the pilot exposed a clear disparity in AI literacy between different educational tracks. Teens in general/academic secondary education tended to score higher on AI knowledge and skills than those in vocational programs. By contrast, factors like age or gender made little difference in the results. This suggests that the school environment and curriculum (what the report calls the educational pathway) play a big role in how much AI understanding a young person picks up. Vocational students – often from more disadvantaged backgrounds – may simply have had fewer opportunities or contexts to learn about AI, whether at school or at home. It’s a reminder that access to AI literacy is uneven, and special attention may be needed to support learners in vocational tracks who are lagging behind their peers.

On a positive note, girls and boys in Flanders performed similarly on the AI literacy scale. This implies that, at least in this sample, the oft-feared digital gender divide might be narrowing when it comes to AI know-how. The more glaring gap is along socio-educational lines.

Another takeaway from the pilot and the preceding interviews is validation of the project’s user-centered approach. The integrated skill use seen in the data affirms what young participants had been telling the researchers all along: they don’t think of “AI skills” in isolation. A single task – say, using an AI-powered app to edit a photo – blends technical operation, creative tweaking, and a bit of ethical sense (is this edit truthful or fair?). The new four-dimensional yAILS model recognizes this blend.

Looking Ahead: Targeted Interventions for Vulnerable Youth

With a validated scale in hand and rich data on where young people stand, the PRODIGI project is now turning insight into action. The next phase will apply these refined yAILS dimensions in tailored interventions among the very groups who showed the greatest needs. In practice, this means developing AI literacy workshops and educational programmes for specific vulnerable youth populations: youngsters with a migration background and those in vocational education. These were core target groups for PRODIGI from the start (e.g. unaccompanied refugee minors in Belgium and disadvantaged vocational students in Portugal), and the pilot results reinforced that they require focused support.

The upcoming interventions will likely leverage the four newly defined skill clusters to create a balanced curriculum. For example, a training session for refugee youth might blend technical and operational guidance (how to safely use AI tools for language learning or job hunting) with content creation exercises and discussions about AI ethics in everyday life. For vocational students, a programme might integrate information navigation skills (like spotting AI-driven misinformation in social media) with hands-on creative projects using AI, all wrapped in an accessible, practice-oriented approach. By clustering activities around the technical, creative, informational, and ethical dimensions identified, educators can ensure they cover the full spectrum of AI literacy in a way that feels relevant and holistic.

Perhaps just as importantly, PRODIGI’s intervention design will continue to be a co-designed effort. The pilot and the preparatory research underscored that listening to youth voices is crucial: what good is an AI literacy lesson if it doesn’t resonate with the learners?

In sum, the PRODIGI project’s Flanders pilot has done more than validate a survey instrument, as the journey from a pilot survey in Flanders to on-the-ground workshops across Europe exemplifies how evidence-based adaptation can drive the PRODIGI mission forward, ensuring no young person is left behind in the age of algorithms.

Source:

Azadi, T., d’Haenens, L., De Nolf, A., Ponte, C., Luna, E., Tomczyk, Ł., Donoso, V., Torres da Silva, M., Batista, S., & Żegleń, M.(2025). Piloting And Validating Tailored Pre- and Post-Intervention Assessments for PRODIGI Target Groups. PRODIGI, KU Leuven.