Should we open the Black Box? Investigating Transparency in K-12 Machine Learning Education

Ferraioli, Valentina; Lehner, Lukas

doi:10.1145/3801749.3801788

Record link:

http://hdl.handle.net/20.500.12708/230495

Title:

Should we open the Black Box? Investigating Transparency in K-12 Machine Learning Education

Citation:

Ferraioli, V., & Lehner, L. (2026). Should we open the Black Box? Investigating Transparency in K-12 Machine Learning Education. In N. Bergner, T. Michaeli, & A. Brodnik (Eds.), WiPSCE ’26 : Proceedings of the 20th WiPSCE Conference on Primary and Secondary Computing Education Research. The Association for Computing Machinery (ACM). https://doi.org/10.1145/3801749.3801788

Publisher DOI:

10.1145/3801749.3801788

Publication Type:

Inproceedings - Abstract Book Contribution

Language:

English

Authors:

Ferraioli, Valentina
Lehner, Lukas

Organisational Unit:

E194-01 - Forschungsbereich Software Engineering

Published in:

WiPSCE '26 : Proceedings of the 20th WiPSCE Conference on Primary and Secondary Computing Education Research

ISBN:

979-8-4007-2502-9

DOI of the book:

10.1145/3801749

Date (published):

20-Aug-2026

Event name:

20th WiPSCE Conference on Primary and Secondary Computing Education Research (WiPSCE '26)

Event date:

11-Mar-2026 - 13-Mar-2026

Event place:

Aachen, Germany

Number of Pages:

Publisher:

The Association for Computing Machinery (ACM), New York, USA

Peer reviewed:

Yes

Keywords:

AI literacy; machine learning; K-12; unplugged activities

Abstract:

The “black box” nature of machine learning (ML) poses a significant pedagogical challenge. To address this, we present a study investigating the impact of “glass-boxing” on students’ understanding of ML concepts. The design uses an unplugged intervention where pipeline transparency is manipulated across three conditions covering data preparation, modelling, and evaluation. This is implemented using variations of the unplugged learning materials. We report on lessons learned from a pilot implementation (N = 66) of the study design and our planned changes.

Project (external):

Wiener Wissenschafts-, Forschungs- und Technologiefonds (WWTF)

Research Areas:

Visual Computing and Human-Centered Technology: 50%
Logic and Computation: 50%

Science Branch:

1020 - Informatik: 50%
5030 - Erziehungswissenschaften: 50%

License:

CC BY 4.0