An XAI Social Media Platform for Teaching K-12 Students AI-Driven Profiling, Clustering, and Engagement-Based Recommending

TL;DR

This paper addresses the need to teach K-12 students how AI-powered social media platforms collect data, build profiles, track engagement, and deliver recommendations. It introduces Somekone, a browser-based explainable AI education tool with an Instagram-like interface that provides real-time explanations and classroom-wide visualizations of data-driven mechanisms. In a pilot with $n=209$ learners across $12$ schools, latent profile analysis identified $k=3$ usage profiles and sequence analysis revealed distinct interaction trajectories, illustrating how students navigate data and recommendations. The work advances AI literacy and data agency in schooling, supports responsible understanding of personalization and privacy, and provides rich data for learning analytics research, while noting limitations and opportunities for stronger ethics integration and scalable teacher materials.

Abstract

This paper, submitted to the special track on resources for teaching AI in K-12, presents an explainable AI (XAI) education tool designed for K-12 classrooms, particularly for students in grades 4-9. The tool was designed for interventions on the fundamental processes behind social media platforms, focusing on four AI- and data-driven core concepts: data collection, user profiling, engagement metrics, and recommendation algorithms. An Instagram-like interface and a monitoring tool for explaining the data-driven processes make these complex ideas accessible and engaging for young learners. The tool provides hands-on experiments and real-time visualizations, illustrating how user actions influence both their personal experience on the platform and the experience of others. This approach seeks to enhance learners' data agency, AI literacy, and sensitivity to AI ethics. The paper includes a case example from 12 two-hour test sessions involving 209 children, using learning analytics to demonstrate how they navigated their social media feeds and the browsing patterns that emerged.

Figures (8)

• Figure 1: Browsing an image feed on Somekone.
• Figure 2: Somekone users can work in pairs, where one user browses the image feed on a mobile device and another connects a second device to that feed to analyze a live view of data collection (Fig. \ref{['fig:app_dataview']}). At the same time, the whole classroom's most engaged images are shown on the classroom projector (Fig. \ref{['fig:cls_engaged']}).
• Figure 3: A paired user device can show a real-time view of another user's profile forming as they browse the feed, with a breakdown that explains the profile (Fig. \ref{['fig:app_profile']}). The teacher's view shown on the classroom projector has dozens of visualization options into the classroom social network (Fig. \ref{['fig:cls_social_network']})
• Figure 4: The recommendations view of Somekone shows, in advance, the next content recommended to the paired user, with a detailed breakdown for each recommendation (Fig. \ref{['fig:app_recommendations']}). The classroom view can provide different visualizations of the clusters that serve as a basis for recommendations (Fig. \ref{['fig:cls_clustering']}.)
• Figure 5: Somekone provides a view of each user's "bubble," or the pool from which the user's recommendations are most likely drawn (Fig. \ref{['fig:app_recspace']}, \ref{['fig:app_recspace2']}). User can experiment on different algorithms and approaches to making recommendations (Fig. \ref{['fig:app_changingrecs']}).