AI from concrete to abstract: demystifying artificial intelligence to the general public

Researchers present AIcon2abs, a methodology combining visual programming with weightless neural networks to teach artificial intelligence concepts to general audiences and children. The approach demystifies AI through hands-on learning activities that integrate training and classification directly into programming blocks, making the distinction between learning and conventional programs more transparent.

The AIcon2abs methodology addresses a critical gap in AI literacy among the general population. As artificial intelligence systems increasingly influence decisions across healthcare, finance, and governance, public understanding of how these systems function has become essential for informed democratic participation. This research recognizes that traditional AI education often treats machine learning as a black-box external module, reinforcing misconceptions about how intelligence works in practice.

The pedagogical innovation lies in integrating neural network training and classification as native programming constructs rather than separate modules. This design choice leverages WiSARD weightless neural networks, which offer inherent simplicity and transparency compared to deep learning approaches. By enabling learners to observe internal processes directly, the methodology transforms passive consumption of AI technology into active comprehension of underlying mechanisms.

From an educational technology perspective, this approach has broad implications. Demystifying AI through concrete programming activities could accelerate adoption of computational thinking in schools while building a more informed citizenry. Teachers and curriculum developers gain a practical toolkit for teaching AI literacy without requiring advanced mathematical backgrounds from students.

The work's impact extends beyond education into civic participation. When general populations understand how learning machines differ from traditional software, they become more effective stakeholders in policy discussions around AI deployment, bias, and regulation. This foundation of practical knowledge supports more sophisticated conversations about algorithmic accountability and ethical AI implementation in society.

• →AIcon2abs integrates AI training and classification as native programming blocks rather than external modules, improving conceptual understanding.
• →Weightless neural networks enable transparent visualization of learning processes, making AI mechanisms accessible to non-experts and children.
• →The methodology bridges critical gaps in public AI literacy needed for informed participation in technology policy and deployment decisions.
• →Practical hands-on learning demonstrates the fundamental difference between learning-capable and conventional programs through direct observation.
• →Simplified AI education architecture reduces barriers to computational thinking adoption in schools and general populations.