← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

Endless Swipes and Recommendations: The Impact of Short-Form Video Platforms on Context-Switching and Children’s Working Memory

Cheryl Siy , Kihoon Jung , Seokwoo Song , Kwan Hong Lee , John Kim

DOI PDF Program page Page JSON

This is a solid CHI-style controlled study with a clear, timely question and a useful null result: platform mechanics increased switching, but the memory measures did not move in the expected way. The contribution is strongest as an empirical finding about immediate effects in a specific child population, not as a broad claim about all short-form video use.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

This paper is best read as a careful, timely empirical contribution rather than a sweeping theory paper. Its main value is that it isolates two salient short-form video mechanics—swipe-based interaction and recommendation-driven content—and tests them in a controlled between-subjects experiment with children, using standard digit-span measures for short-term and working memory. That design gives the authors a credible basis for saying that the platform features changed switching behavior, while the memory outcomes themselves did not show the intuitive direct degradation one might expect from rapid context switching. In CHI terms, that is a meaningful departure from common sense because it complicates a simple causal story that more switching automatically means worse memory. The novelty is also reasonably clear: the paper explicitly positions itself as addressing an underexplored gap, especially for young children, and it does so with a concrete experimental setup rather than a speculative discussion. At the same time, the scope is narrow and should be kept that way. The evidence is immediate, lab-based, and tied to a 10-minute exposure on YouTube Shorts, so it supports claims about short-term experimental effects, not long-term developmental consequences or broad platform harms. The authors themselves acknowledge important limits around cross-sectional inference, practice effects, grade-level variability, proprietary recommendation logic, and hardware/screen differences. So the strongest expert reading is: a well-executed controlled study with a useful null result and a modest but real contribution to child-focused HCI and media-effects discussions, but not a basis for strong generalization beyond the tested setting.

What Changed

Canon before

Prior CHI work on children, short-form video, and cognition had not directly established whether platform mechanics like swipe-driven switching and recommendation-driven content selection alter working memory or short-term memory under controlled exposure.

Departure from common sense

The paper challenges the intuitive expectation that rapid context switching from short-form video should directly degrade children’s memory performance: the reported feature combinations and baseline comparisons were not associated with changes in STM or WM, even though switching behavior itself increased.

Actual novelty

The paper’s novelty is in directly testing short-form video platform mechanics in young children and linking swipe/recommendation features to switching behavior while measuring STM and WM. The authors explicitly frame this as an area that had not been evaluated, especially in young children.

Evidence

A between-subjects experiment with N=180 children used digit span forward/backward tasks after controlled YouTube Shorts exposure. The design varied swipe interaction and content source in a 2x2 factorial setup with long-form baselines. Results reported no STM/WM differences across feature combinations or baselines, but did find increased switching from swipe and recommendation conditions and marginal WM associations with higher switching.

Limits

Method limits

The evidence comes from a controlled, short-duration experiment rather than a longitudinal or naturalistic deployment. The study measures immediate effects after a 10-minute exposure and therefore cannot establish durable cognitive impact over time.

Deployment limits

Findings are specific to the tested platform context, YouTube Shorts, the selected age group, and the particular experimental implementation of swipe and recommendation conditions. Generalization to other short-form video platforms, older users, or longer-term everyday use is limited.

Boundary conditions

The reported effects are bounded by a 10-minute intervention, grade-level variability in the sample, and the use of digit-span tasks as the cognitive outcome. The paper also notes constraints from proprietary recommendation systems and hardware/screen differences.

Manual check: No manual check required for the extracted claims, but the exact wording of the limitation evidence is constrained by the provided spans and should be rechecked against the full text if publication decisions depend on it.

Position in field

This work sits at the intersection of HCI, child-computer interaction, and media effects research. It contributes an experimentally grounded CHI result on how short-form video mechanics relate to switching behavior and memory outcomes in children, with a notable null result on STM/WM despite increased switching.

Abstract

Short-form video (SFV) platforms are increasingly popular, yet the rapid context switching and their potential effects on children’s cognitive functions are not well understood. In this work, we conducted a between-subjects experiment ($N = 180$) to examine how YouTube Shorts affects young children’s short-term memory (STM) and working memory (WM), measured using the forward and backward digit span tasks. The study focused on two core platform features of SFV: the easy-to-use swipe interface and the recommendation system. Using a 2$\times$2 factorial design, we compared four SFV group conditions that varied by interaction mode and content source, complemented by two long-form video baseline conditions (one with constant context switching and one without). Our results show that the feature combinations and the baseline comparisons were not associated with changes in STM or WM. However, swipe interaction increased video switching, while recommendation-based content increased category switching. The higher combined levels of video and category switching across participants were associated with marginal effects on working memory performance, while STM remained unaffected.