Executive Summary

This White Paper was commissioned by the Service national de la Jeunesse (SNJ), a public institution governed by the Ministère de l'Éducation nationale, de l'Enfance et de la Jeunesse (MENJE), and prepared by researchers from the University of Luxembourg — specifically ULIDE (University of Luxembourg Institute for Digital Ethics), LUCET (Luxembourg Centre for Educational Testing) and CCY (Centre for Childhood and Youth Research). It addresses the challenges and risks posed by artificial intelligence (AI) in relation to internet use among children and adolescents. We focus in particular on generative AI systems, AI-driven recommendation infrastructures, virtual companions, and synthetic content.

This paper emphasises that AI is no longer a separate technology; it is an invisible layer of digital mediation that shapes how young people access information, communicate, learn, and form opinions. Children and adolescents are increasingly exposed to AI systems — often unintentionally — through social media feeds, gaming environments, search tools, and educational applications. This exposure occurs at stages of cognitive, emotional, and social development that make young users particularly susceptible to influence, manipulation, and dependency.

Our analysis identifies three overarching challenges for policy and practice:

1. AI exacerbates structural power imbalances between young users and commercial platforms that prioritise engagement, personalisation, and data extraction while remaining largely opaque.
2. AI-generated outputs tend to acquire epistemic authority: they are perceived as neutral or objective even though they may reproduce bias, misinformation, or strategic framing.
3. Privacy and data protection risks are amplified by continuous personalisation, emotional profiling, and the processing of sensitive behavioural data.

In educational settings, AI offers significant potential for personalised learning, feedback, and accessibility, but also poses systemic risks such as the erosion of critical thinking, cognitive offloading, unequal learning conditions, and challenges to assessment integrity. This White Paper emphasises that the central question is not whether AI will be used by children and adolescents, but how its use is to be shaped, guided, and governed.

A dedicated section analyses BEE SECURE’s existing activities and instruments, including awareness campaigns, educational materials, training programmes, and counselling services. Our assessment highlights BEE SECURE’s strong positioning, trusted role, and well-established prevention infrastructure as a solid foundation for addressing AI-related challenges. Building on this, the White Paper formulates concrete, forward-looking recommendations on how BEE SECURE can systematically expand its mandate to cover AI-specific risks — including generative systems, synthetic identities, and algorithmic influence — through targeted adaptations of its educational content, reporting mechanisms, counselling services, and internal expertise.

For Luxembourg, the report formulates six strategic priorities:

1. integrating AI literacy into existing prevention and education programmes;
2. equipping teachers, parents, and youth professionals with practical guidance;
3. strengthening reporting and counselling mechanisms for AI-related harms;
4. institutionalising youth participation;
5. embedding ethical standards and monitoring capacities;
6. and anchoring policy development in continuous scientific evaluation.

Overall, this White Paper frames AI and youth as core issues of digital safety, education, and democratic resilience. It positions BEE SECURE as a key actor in this evolving landscape and underscores the need for early, coordinated, and evidence-based public action to ensure that AI supports — rather than undermines — the autonomy, well-being, and learning capacities of children and adolescents.